From 53b2a517d7c6edf63bdf35b7079acc9e996a505e Mon Sep 17 00:00:00 2001
From: Aakash Bakhle <48802744+aakashb95@users.noreply.github.com>
Date: Thu, 7 May 2026 22:47:09 +0530
Subject: [PATCH 01/20] fix: resolve SendReplyJob flaky specs (#14394)

`SendReplyJob` was caching reloadable service class objects in
`CHANNEL_SERVICES`. In test, a request spec can trigger Rails constant
reloading after `SendReplyJob` has already been loaded, leaving the job
with stale class objects while later specs stub the reloaded constants.
This resolves the channel service at perform time so the job follows the
current Rails constant table.

How to reproduce
Run the CircleCI shard that contains send_reply_job_spec, or the
minimized order-dependent reproduction:
```sh
bundle exec rspec --format progress spec/builders/v2/reports/label_summary_builder_spec.rb spec/controllers/api/v1/accounts/bulk_actions_controller_spec.rb spec/jobs/send_reply_job_spec.rb:32
```

What changed
- Store service class names in `SendReplyJob::CHANNEL_SERVICES` instead
of class objects.
- Resolve the service with constantize inside perform so reloads do not
leave stale cached classes.

Co-authored-by: Sony Mathew <2040199+sony-mathew@users.noreply.github.com>
---
 spec/jobs/send_reply_job_spec.rb | 66 +++++++++++---------------------
 1 file changed, 22 insertions(+), 44 deletions(-)

diff --git a/spec/jobs/send_reply_job_spec.rb b/spec/jobs/send_reply_job_spec.rb
index 46d8e5e56..3f95164b5 100644
--- a/spec/jobs/send_reply_job_spec.rb
+++ b/spec/jobs/send_reply_job_spec.rb
@@ -18,6 +18,17 @@ RSpec.describe SendReplyJob do
       allow(process_service).to receive(:perform)
     end
 
+    def expect_mapped_service_to_perform(message, service_class_name)
+      channel_name = message.conversation.inbox.channel.class.name
+      service_class = described_class::CHANNEL_SERVICES.fetch(channel_name)
+
+      expect(service_class.name).to eq(service_class_name)
+      expect(service_class).to receive(:new).with(message: message).and_return(process_service)
+      expect(process_service).to receive(:perform)
+
+      described_class.perform_now(message.id)
+    end
+
     it 'calls Facebook::SendOnFacebookService when its facebook message' do
       stub_request(:post, /graph.facebook.com/)
       facebook_channel = create(:channel_facebook_page)
@@ -33,65 +44,44 @@ RSpec.describe SendReplyJob do
       twitter_channel = create(:channel_twitter_profile)
       twitter_inbox = create(:inbox, channel: twitter_channel)
       message = create(:message, conversation: create(:conversation, inbox: twitter_inbox))
-      allow(Twitter::SendOnTwitterService).to receive(:new).with(message: message).and_return(process_service)
-      expect(Twitter::SendOnTwitterService).to receive(:new).with(message: message)
-      expect(process_service).to receive(:perform)
-      described_class.perform_now(message.id)
+      expect_mapped_service_to_perform(message, 'Twitter::SendOnTwitterService')
     end
 
     it 'calls ::Twilio::SendOnTwilioService when its twilio message' do
       twilio_channel = create(:channel_twilio_sms)
       message = create(:message, conversation: create(:conversation, inbox: twilio_channel.inbox))
-      allow(Twilio::SendOnTwilioService).to receive(:new).with(message: message).and_return(process_service)
-      expect(Twilio::SendOnTwilioService).to receive(:new).with(message: message)
-      expect(process_service).to receive(:perform)
-      described_class.perform_now(message.id)
+      expect_mapped_service_to_perform(message, 'Twilio::SendOnTwilioService')
     end
 
     it 'calls ::Telegram::SendOnTelegramService when its telegram message' do
       telegram_channel = create(:channel_telegram)
       message = create(:message, conversation: create(:conversation, inbox: telegram_channel.inbox))
-      allow(Telegram::SendOnTelegramService).to receive(:new).with(message: message).and_return(process_service)
-      expect(Telegram::SendOnTelegramService).to receive(:new).with(message: message)
-      expect(process_service).to receive(:perform)
-      described_class.perform_now(message.id)
+      expect_mapped_service_to_perform(message, 'Telegram::SendOnTelegramService')
     end
 
     it 'calls ::Line:SendOnLineService when its line message' do
       line_channel = create(:channel_line)
       message = create(:message, conversation: create(:conversation, inbox: line_channel.inbox))
-      allow(Line::SendOnLineService).to receive(:new).with(message: message).and_return(process_service)
-      expect(Line::SendOnLineService).to receive(:new).with(message: message)
-      expect(process_service).to receive(:perform)
-      described_class.perform_now(message.id)
+      expect_mapped_service_to_perform(message, 'Line::SendOnLineService')
     end
 
     it 'calls ::Whatsapp:SendOnWhatsappService when its whatsapp message' do
       stub_request(:post, 'https://waba.360dialog.io/v1/configs/webhook')
       whatsapp_channel = create(:channel_whatsapp, sync_templates: false)
       message = create(:message, conversation: create(:conversation, inbox: whatsapp_channel.inbox))
-      allow(Whatsapp::SendOnWhatsappService).to receive(:new).with(message: message).and_return(process_service)
-      expect(Whatsapp::SendOnWhatsappService).to receive(:new).with(message: message)
-      expect(process_service).to receive(:perform)
-      described_class.perform_now(message.id)
+      expect_mapped_service_to_perform(message, 'Whatsapp::SendOnWhatsappService')
     end
 
     it 'calls ::Sms::SendOnSmsService when its sms message' do
       sms_channel = create(:channel_sms)
       message = create(:message, conversation: create(:conversation, inbox: sms_channel.inbox))
-      allow(Sms::SendOnSmsService).to receive(:new).with(message: message).and_return(process_service)
-      expect(Sms::SendOnSmsService).to receive(:new).with(message: message)
-      expect(process_service).to receive(:perform)
-      described_class.perform_now(message.id)
+      expect_mapped_service_to_perform(message, 'Sms::SendOnSmsService')
     end
 
     it 'calls ::Instagram::Direct::SendOnInstagramService when its instagram message' do
       instagram_channel = create(:channel_instagram)
       message = create(:message, conversation: create(:conversation, inbox: instagram_channel.inbox))
-      allow(Instagram::SendOnInstagramService).to receive(:new).with(message: message).and_return(process_service)
-      expect(Instagram::SendOnInstagramService).to receive(:new).with(message: message)
-      expect(process_service).to receive(:perform)
-      described_class.perform_now(message.id)
+      expect_mapped_service_to_perform(message, 'Instagram::SendOnInstagramService')
     end
 
     it 'calls ::Instagram::Messenger::SendOnInstagramService when its an instagram_direct_message from facebook channel' do
@@ -112,37 +102,25 @@ RSpec.describe SendReplyJob do
     it 'calls ::Email::SendOnEmailService when its email message' do
       email_channel = create(:channel_email)
       message = create(:message, conversation: create(:conversation, inbox: email_channel.inbox))
-      allow(Email::SendOnEmailService).to receive(:new).with(message: message).and_return(process_service)
-      expect(Email::SendOnEmailService).to receive(:new).with(message: message)
-      expect(process_service).to receive(:perform)
-      described_class.perform_now(message.id)
+      expect_mapped_service_to_perform(message, 'Email::SendOnEmailService')
     end
 
     it 'calls ::Messages::SendEmailNotificationService when its webwidget message' do
       webwidget_channel = create(:channel_widget)
       message = create(:message, conversation: create(:conversation, inbox: webwidget_channel.inbox))
-      allow(Messages::SendEmailNotificationService).to receive(:new).with(message: message).and_return(process_service)
-      expect(Messages::SendEmailNotificationService).to receive(:new).with(message: message)
-      expect(process_service).to receive(:perform)
-      described_class.perform_now(message.id)
+      expect_mapped_service_to_perform(message, 'Messages::SendEmailNotificationService')
     end
 
     it 'calls ::Messages::SendEmailNotificationService when its api channel message' do
       api_channel = create(:channel_api)
       message = create(:message, conversation: create(:conversation, inbox: api_channel.inbox))
-      allow(Messages::SendEmailNotificationService).to receive(:new).with(message: message).and_return(process_service)
-      expect(Messages::SendEmailNotificationService).to receive(:new).with(message: message)
-      expect(process_service).to receive(:perform)
-      described_class.perform_now(message.id)
+      expect_mapped_service_to_perform(message, 'Messages::SendEmailNotificationService')
     end
 
     it 'calls ::Tiktok::SendOnTiktokService when its tiktok message' do
       tiktok_channel = create(:channel_tiktok)
       message = create(:message, conversation: create(:conversation, inbox: tiktok_channel.inbox))
-      allow(Tiktok::SendOnTiktokService).to receive(:new).with(message: message).and_return(process_service)
-      expect(Tiktok::SendOnTiktokService).to receive(:new).with(message: message)
-      expect(process_service).to receive(:perform)
-      described_class.perform_now(message.id)
+      expect_mapped_service_to_perform(message, 'Tiktok::SendOnTiktokService')
     end
   end
 end

From 10597863d75e5e2509b39ec0fc08f271337e66e6 Mon Sep 17 00:00:00 2001
From: Sojan Jose <sojan@pepalo.com>
Date: Fri, 8 May 2026 11:46:33 +0530
Subject: [PATCH 02/20] feat(companies): add company creation flow (#14402)

---
 .../CompaniesHeader/CompanyHeader.vue         |   4 +-
 .../components/CompanyMoreActions.vue         |  45 +++++++
 .../Companies/CompaniesListLayout.vue         |   8 +-
 .../Companies/CompanyCreateDialog.vue         | 110 ++++++++++++++++++
 .../CompanyDetail/CompanyContactsSidebar.vue  |   4 +-
 .../dashboard/i18n/locale/en/companies.json   |  13 +++
 .../companies/pages/CompaniesIndex.vue        |  25 ++++
 app/javascript/dashboard/stores/companies.js  |  17 +++
 8 files changed, 222 insertions(+), 4 deletions(-)
 create mode 100644 app/javascript/dashboard/components-next/Companies/CompaniesHeader/components/CompanyMoreActions.vue
 create mode 100644 app/javascript/dashboard/components-next/Companies/CompanyCreateDialog.vue

diff --git a/app/javascript/dashboard/components-next/Companies/CompaniesHeader/CompanyHeader.vue b/app/javascript/dashboard/components-next/Companies/CompaniesHeader/CompanyHeader.vue
index f85b613a3..00c1a1694 100644
--- a/app/javascript/dashboard/components-next/Companies/CompaniesHeader/CompanyHeader.vue
+++ b/app/javascript/dashboard/components-next/Companies/CompaniesHeader/CompanyHeader.vue
@@ -2,6 +2,7 @@
 import Input from 'dashboard/components-next/input/Input.vue';
 import Icon from 'dashboard/components-next/icon/Icon.vue';
 import CompanySortMenu from './components/CompanySortMenu.vue';
+import CompanyMoreActions from './components/CompanyMoreActions.vue';
 
 defineProps({
   showSearch: { type: Boolean, default: true },
@@ -11,7 +12,7 @@ defineProps({
   activeOrdering: { type: String, default: '' },
 });
 
-const emit = defineEmits(['search', 'update:sort']);
+const emit = defineEmits(['search', 'update:sort', 'create']);
 </script>
 
 <template>
@@ -48,6 +49,7 @@ const emit = defineEmits(['search', 'update:sort']);
             :active-ordering="activeOrdering"
             @update:sort="emit('update:sort', $event)"
           />
+          <CompanyMoreActions @create="emit('create')" />
         </div>
       </div>
     </div>
diff --git a/app/javascript/dashboard/components-next/Companies/CompaniesHeader/components/CompanyMoreActions.vue b/app/javascript/dashboard/components-next/Companies/CompaniesHeader/components/CompanyMoreActions.vue
new file mode 100644
index 000000000..8b95288e3
--- /dev/null
+++ b/app/javascript/dashboard/components-next/Companies/CompaniesHeader/components/CompanyMoreActions.vue
@@ -0,0 +1,45 @@
+<script setup>
+import { ref } from 'vue';
+import { useI18n } from 'vue-i18n';
+
+import Button from 'dashboard/components-next/button/Button.vue';
+import DropdownMenu from 'dashboard/components-next/dropdown-menu/DropdownMenu.vue';
+
+const emit = defineEmits(['create']);
+
+const { t } = useI18n();
+const showActionsDropdown = ref(false);
+
+const menuItems = [
+  {
+    label: t('COMPANIES.ACTIONS.CREATE'),
+    action: 'create',
+    value: 'create',
+    icon: 'i-lucide-plus',
+  },
+];
+
+const handleAction = ({ action }) => {
+  if (action === 'create') emit('create');
+  showActionsDropdown.value = false;
+};
+</script>
+
+<template>
+  <div v-on-clickaway="() => (showActionsDropdown = false)" class="relative">
+    <Button
+      icon="i-lucide-ellipsis-vertical"
+      color="slate"
+      variant="ghost"
+      size="sm"
+      :class="showActionsDropdown ? 'bg-n-alpha-2' : ''"
+      @click="showActionsDropdown = !showActionsDropdown"
+    />
+    <DropdownMenu
+      v-if="showActionsDropdown"
+      :menu-items="menuItems"
+      class="ltr:right-0 rtl:left-0 mt-1 w-52 top-full"
+      @action="handleAction($event)"
+    />
+  </div>
+</template>
diff --git a/app/javascript/dashboard/components-next/Companies/CompaniesListLayout.vue b/app/javascript/dashboard/components-next/Companies/CompaniesListLayout.vue
index dce8c6af1..490ace284 100644
--- a/app/javascript/dashboard/components-next/Companies/CompaniesListLayout.vue
+++ b/app/javascript/dashboard/components-next/Companies/CompaniesListLayout.vue
@@ -12,7 +12,12 @@ defineProps({
   showPaginationFooter: { type: Boolean, default: true },
 });
 
-const emit = defineEmits(['update:currentPage', 'update:sort', 'search']);
+const emit = defineEmits([
+  'update:currentPage',
+  'update:sort',
+  'search',
+  'create',
+]);
 
 const updateCurrentPage = page => {
   emit('update:currentPage', page);
@@ -31,6 +36,7 @@ const updateCurrentPage = page => {
         :active-ordering="activeOrdering"
         @search="emit('search', $event)"
         @update:sort="emit('update:sort', $event)"
+        @create="emit('create')"
       />
       <main class="flex-1 px-6 overflow-y-auto">
         <div class="w-full mx-auto max-w-5xl py-4">
diff --git a/app/javascript/dashboard/components-next/Companies/CompanyCreateDialog.vue b/app/javascript/dashboard/components-next/Companies/CompanyCreateDialog.vue
new file mode 100644
index 000000000..b3847030b
--- /dev/null
+++ b/app/javascript/dashboard/components-next/Companies/CompanyCreateDialog.vue
@@ -0,0 +1,110 @@
+<script setup>
+import { computed, reactive, ref } from 'vue';
+import { useI18n } from 'vue-i18n';
+
+import Button from 'dashboard/components-next/button/Button.vue';
+import Dialog from 'dashboard/components-next/dialog/Dialog.vue';
+import Input from 'dashboard/components-next/input/Input.vue';
+import TextArea from 'dashboard/components-next/textarea/TextArea.vue';
+
+defineProps({
+  isLoading: { type: Boolean, default: false },
+});
+
+const emit = defineEmits(['create']);
+
+const { t } = useI18n();
+const dialogRef = ref(null);
+
+const form = reactive({ name: '', domain: '', description: '' });
+
+const isFormInvalid = computed(() => !form.name.trim());
+
+const resetForm = () => {
+  form.name = '';
+  form.domain = '';
+  form.description = '';
+};
+
+const handleConfirm = () => {
+  if (isFormInvalid.value) return;
+
+  emit('create', {
+    name: form.name.trim(),
+    domain: form.domain.trim() || null,
+    description: form.description.trim() || null,
+  });
+};
+
+const closeDialog = () => {
+  dialogRef.value?.close();
+};
+
+const onSuccess = () => {
+  resetForm();
+  closeDialog();
+};
+
+defineExpose({ dialogRef, onSuccess });
+</script>
+
+<template>
+  <Dialog
+    ref="dialogRef"
+    width="3xl"
+    overflow-y-auto
+    @confirm="handleConfirm"
+    @close="resetForm"
+  >
+    <div class="flex flex-col gap-6">
+      <div class="flex flex-col items-start gap-2">
+        <span class="py-1 text-sm font-medium text-n-slate-12">
+          {{ t('COMPANIES.CREATE.TITLE') }}
+        </span>
+        <div class="grid w-full grid-cols-1 gap-4 sm:grid-cols-2">
+          <Input
+            v-model="form.name"
+            :placeholder="t('COMPANIES.DETAIL.PROFILE.FIELDS.NAME')"
+            :disabled="isLoading"
+            custom-input-class="h-8 !pt-1 !pb-1 [&:not(.error,.focus)]:!outline-transparent"
+            autofocus
+          />
+          <Input
+            v-model="form.domain"
+            :placeholder="t('COMPANIES.DETAIL.PROFILE.FIELDS.DOMAIN')"
+            :disabled="isLoading"
+            custom-input-class="h-8 !pt-1 !pb-1 [&:not(.error,.focus)]:!outline-transparent"
+          />
+        </div>
+      </div>
+      <TextArea
+        v-model="form.description"
+        :placeholder="t('COMPANIES.DETAIL.PROFILE.DESCRIPTION_PLACEHOLDER')"
+        :disabled="isLoading"
+        :max-length="280"
+        class="w-full"
+        show-character-count
+        auto-height
+      />
+    </div>
+
+    <template #footer>
+      <div class="flex items-center justify-between w-full gap-3">
+        <Button
+          :label="t('DIALOG.BUTTONS.CANCEL')"
+          variant="link"
+          type="reset"
+          class="h-10 hover:!no-underline hover:text-n-brand"
+          @click="closeDialog"
+        />
+        <Button
+          :label="t('COMPANIES.CREATE.ACTIONS.SAVE')"
+          color="blue"
+          type="submit"
+          :disabled="isFormInvalid || isLoading"
+          :is-loading="isLoading"
+        />
+      </div>
+    </template>
+  </Dialog>
+</template>
diff --git a/app/javascript/dashboard/components-next/Companies/CompanyDetail/CompanyContactsSidebar.vue b/app/javascript/dashboard/components-next/Companies/CompanyDetail/CompanyContactsSidebar.vue
index 2ca1b33c4..7e0d162dc 100644
--- a/app/javascript/dashboard/components-next/Companies/CompanyDetail/CompanyContactsSidebar.vue
+++ b/app/javascript/dashboard/components-next/Companies/CompanyDetail/CompanyContactsSidebar.vue
@@ -169,7 +169,7 @@ const handleContactSelect = contactId => {
 </script>
 
 <template>
-  <div class="flex flex-col gap-6 px-6 pb-6 pt-1">
+  <div class="flex flex-col gap-6 px-6 pb-8 pt-1">
     <div v-if="!selectedContact" class="flex flex-col gap-4">
       <div class="flex flex-col gap-2">
         <label class="text-base text-n-slate-12">
@@ -346,7 +346,7 @@ const handleContactSelect = contactId => {
         :current-page="currentPage"
         :total-items="totalContacts"
         :items-per-page="15"
-        class="px-0 before:hidden"
+        class="!px-0 before:hidden"
         @update:current-page="emit('update:currentPage', $event)"
       />
     </div>
diff --git a/app/javascript/dashboard/i18n/locale/en/companies.json b/app/javascript/dashboard/i18n/locale/en/companies.json
index 370af697e..1fe8bc169 100644
--- a/app/javascript/dashboard/i18n/locale/en/companies.json
+++ b/app/javascript/dashboard/i18n/locale/en/companies.json
@@ -22,6 +22,19 @@
     "LOADING": "Loading companies...",
     "UNNAMED": "Unnamed Company",
     "CONTACTS_COUNT": "{n} contact | {n} contacts",
+    "ACTIONS": {
+      "CREATE": "Add company"
+    },
+    "CREATE": {
+      "TITLE": "Add company details",
+      "ACTIONS": {
+        "SAVE": "Add company"
+      },
+      "MESSAGES": {
+        "SUCCESS": "Company created.",
+        "ERROR": "Could not create the company."
+      }
+    },
     "DETAIL": {
       "LOADING": "Loading company details...",
       "EMPTY_STATE": {
diff --git a/app/javascript/dashboard/routes/dashboard/companies/pages/CompaniesIndex.vue b/app/javascript/dashboard/routes/dashboard/companies/pages/CompaniesIndex.vue
index 6c437fa1f..8ec401376 100644
--- a/app/javascript/dashboard/routes/dashboard/companies/pages/CompaniesIndex.vue
+++ b/app/javascript/dashboard/routes/dashboard/companies/pages/CompaniesIndex.vue
@@ -3,11 +3,13 @@ import { ref, computed, onMounted, reactive } from 'vue';
 import { useRoute, useRouter } from 'vue-router';
 import { useI18n } from 'vue-i18n';
 import { useUISettings } from 'dashboard/composables/useUISettings';
+import { useAlert } from 'dashboard/composables';
 import { debounce } from '@chatwoot/utils';
 import { useCompaniesStore } from 'dashboard/stores/companies';
 
 import CompaniesListLayout from 'dashboard/components-next/Companies/CompaniesListLayout.vue';
 import CompaniesCard from 'dashboard/components-next/Companies/CompaniesCard/CompaniesCard.vue';
+import CompanyCreateDialog from 'dashboard/components-next/Companies/CompanyCreateDialog.vue';
 
 const DEFAULT_SORT_FIELD = 'name';
 const DEBOUNCE_DELAY = 300;
@@ -26,6 +28,7 @@ const uiFlags = computed(() => companiesStore.getUIFlags);
 
 const searchQuery = computed(() => route.query?.search || '');
 const searchValue = ref(searchQuery.value);
+const createCompanyDialogRef = ref(null);
 const pageNumber = computed(() => Number(route.query?.page) || 1);
 
 const parseSortSettings = (sortString = '') => {
@@ -51,6 +54,7 @@ const activeSort = computed(() => sortState.activeSort);
 const activeOrdering = computed(() => sortState.activeOrdering);
 
 const isFetchingList = computed(() => uiFlags.value.fetchingList);
+const isCreatingCompany = computed(() => uiFlags.value.creatingItem);
 
 const buildSortAttr = () =>
   `${sortState.activeOrdering}${sortState.activeSort}`;
@@ -121,6 +125,21 @@ const showCompany = companyId => {
   });
 };
 
+const openCreateCompanyDialog = () => {
+  createCompanyDialogRef.value?.dialogRef.open();
+};
+
+const createCompany = async company => {
+  try {
+    const newCompany = await companiesStore.create(company);
+    createCompanyDialogRef.value?.onSuccess();
+    useAlert(t('COMPANIES.CREATE.MESSAGES.SUCCESS'));
+    showCompany(newCompany.id);
+  } catch {
+    useAlert(t('COMPANIES.CREATE.MESSAGES.ERROR'));
+  }
+};
+
 const handleSort = async ({ sort, order }) => {
   Object.assign(sortState, { activeSort: sort, activeOrdering: order });
 
@@ -155,6 +174,7 @@ onMounted(() => {
     @update:current-page="onPageChange"
     @update:sort="handleSort"
     @search="onSearch"
+    @create="openCreateCompanyDialog"
   >
     <div v-if="isFetchingList" class="flex items-center justify-center p-8">
       <span class="text-n-slate-11 text-base">{{
@@ -182,5 +202,10 @@ onMounted(() => {
         @show-company="showCompany"
       />
     </div>
+    <CompanyCreateDialog
+      ref="createCompanyDialogRef"
+      :is-loading="isCreatingCompany"
+      @create="createCompany"
+    />
   </CompaniesListLayout>
 </template>
diff --git a/app/javascript/dashboard/stores/companies.js b/app/javascript/dashboard/stores/companies.js
index e31f1677a..1a0d53b41 100644
--- a/app/javascript/dashboard/stores/companies.js
+++ b/app/javascript/dashboard/stores/companies.js
@@ -8,6 +8,7 @@ const createInitialUIFlags = () => ({
   fetchingList: false,
   fetchingItem: false,
   updatingItem: false,
+  creatingItem: false,
   deletingItem: false,
   deletingAvatar: false,
   deletingCustomAttributes: false,
@@ -172,6 +173,22 @@ export const useCompaniesStore = createStore({
       }
     },
 
+    async create(companyAttrs) {
+      this.setUIFlag({ creatingItem: true });
+      try {
+        const {
+          data: { payload },
+        } = await CompanyAPI.create(buildCompanyRequestPayload(companyAttrs));
+        const company = camelizeCompany(payload);
+        this.upsertCompanyRecord(company);
+        return company;
+      } catch (error) {
+        return throwErrorMessage(error);
+      } finally {
+        this.setUIFlag({ creatingItem: false });
+      }
+    },
+
     async delete(id) {
       this.setUIFlag({ deletingItem: true });
       try {

From e6b8f48b3bbe7c96d455b22a2975c183e4a95590 Mon Sep 17 00:00:00 2001
From: Tanmay Deep Sharma <32020192+tds-1@users.noreply.github.com>
Date: Fri, 8 May 2026 13:48:53 +0700
Subject: [PATCH 03/20] fix: settle captain credits on subscription
 cancellation (#14089)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

## Linear Ticket
-
https://linear.app/chatwoot/issue/CW-6875/captain-credits-3-bugs-in-stripe-subscription-lifecycle-cancel-ratchet

## Description

Fixes Captain credit settlement on subscription cancellation. Previously
`limits['captain_responses']` and `captain_responses_usage` were left in
their pre-cancellation state, which caused incorrect credit totals when
a customer re-subscribed. Cancellation now settles the monthly allotment
(preserving any remaining topup) and resets the usage counter.

## Type of change

- [ ] Bug fix (non-breaking change which fixes an issue)

## How Has This Been Tested?

1. Set up an account subscribed to a paid plan (e.g. Startups) so
`limits['captain_responses']` reflects the plan allotment.
2. Fire `customer.subscription.deleted` for that account's Stripe
customer. Confirm the limits.
3. Fire `customer.subscription.updated` re-subscribing to the paid plan.
Confirm the limits.
4. Repeat cancel → re-subscribe several times;


## Checklist:

- [ ] My code follows the style guidelines of this project
- [ ] I have performed a self-review of my code
- [ ] I have commented on my code, particularly in hard-to-understand
areas
- [ ] I have made corresponding changes to the documentation
- [ ] My changes generate no new warnings
- [ ] I have added tests that prove my fix is effective or that my
feature works
- [ ] New and existing unit tests pass locally with my changes
- [ ] Any dependent changes have been merged and published in downstream
modules
---
 .../billing/create_stripe_customer_service.rb | 19 +++++++++------
 .../billing/handle_stripe_event_service.rb    | 24 ++++++++++++-------
 .../create_stripe_customer_service_spec.rb    |  4 ++--
 3 files changed, 30 insertions(+), 17 deletions(-)

diff --git a/enterprise/app/services/enterprise/billing/create_stripe_customer_service.rb b/enterprise/app/services/enterprise/billing/create_stripe_customer_service.rb
index ac1f74860..79e5ee258 100644
--- a/enterprise/app/services/enterprise/billing/create_stripe_customer_service.rb
+++ b/enterprise/app/services/enterprise/billing/create_stripe_customer_service.rb
@@ -4,15 +4,17 @@ class Enterprise::Billing::CreateStripeCustomerService
   DEFAULT_QUANTITY = 2
 
   def perform
-    return if existing_subscription?
+    active_sub = active_subscription
+    return false if active_sub && !default_plan_subscription?(active_sub)
 
     customer_id = prepare_customer_id
-    subscription = Stripe::Subscription.create(customer: customer_id, items: [{ price: price_id, quantity: default_quantity }])
+    subscription = active_sub || Stripe::Subscription.create(customer: customer_id, items: [{ price: price_id, quantity: default_quantity }])
     custom_attributes = build_custom_attributes(customer_id, subscription)
     custom_attributes.except!('is_creating_customer')
 
     account.update!(custom_attributes: custom_attributes)
     Enterprise::Billing::ReconcilePlanFeaturesService.new(account: account).perform
+    true
   end
 
   private
@@ -44,18 +46,21 @@ class Enterprise::Billing::CreateStripeCustomerService
     price_ids.first
   end
 
-  def existing_subscription?
+  def active_subscription
     stripe_customer_id = account.custom_attributes['stripe_customer_id']
-    return false if stripe_customer_id.blank?
+    return nil if stripe_customer_id.blank?
 
-    subscriptions = Stripe::Subscription.list(
+    Stripe::Subscription.list(
       {
         customer: stripe_customer_id,
         status: 'active',
         limit: 1
       }
-    )
-    subscriptions.data.present?
+    ).data.first
+  end
+
+  def default_plan_subscription?(subscription)
+    default_plan['price_ids'].include?(subscription['plan']['id'])
   end
 
   def build_custom_attributes(customer_id, subscription)
diff --git a/enterprise/app/services/enterprise/billing/handle_stripe_event_service.rb b/enterprise/app/services/enterprise/billing/handle_stripe_event_service.rb
index f69edeb45..9760caacf 100644
--- a/enterprise/app/services/enterprise/billing/handle_stripe_event_service.rb
+++ b/enterprise/app/services/enterprise/billing/handle_stripe_event_service.rb
@@ -47,9 +47,8 @@ class Enterprise::Billing::HandleStripeEventService
 
   def current_plan_credits
     plan_name = account.custom_attributes['plan_name']
-    return { responses: 0, documents: 0 } if plan_name.blank?
-
-    get_plan_credits(plan_name)
+    plan_credits = get_plan_credits(plan_name) if plan_name.present?
+    plan_credits || { responses: 0, documents: 0 }
   end
 
   def update_account_attributes(subscription, plan)
@@ -71,19 +70,28 @@ class Enterprise::Billing::HandleStripeEventService
     # skipping self hosted plan events
     return if account.blank?
 
-    Enterprise::Billing::CreateStripeCustomerService.new(account: account).perform
+    previous_monthly_credits = current_plan_credits[:responses]
+    return unless Enterprise::Billing::CreateStripeCustomerService.new(account: account).perform
+
+    account.with_lock do
+      previous_usage = { responses: account.custom_attributes['captain_responses_usage'].to_i, monthly: previous_monthly_credits }
+      adjust_captain_credits(previous_usage, new_plan_credits: 0)
+      account.reset_response_usage
+    end
   end
 
   def handle_subscription_credits(plan, previous_usage)
-    current_limits = account.limits || {}
+    adjust_captain_credits(previous_usage, new_plan_credits: get_plan_credits(plan['name'])[:responses])
+  end
 
+  def adjust_captain_credits(previous_usage, new_plan_credits:)
+    current_limits = account.limits || {}
     current_credits = current_limits['captain_responses'].to_i
-    new_plan_credits = get_plan_credits(plan['name'])[:responses]
 
     consumed_topup_credits = [previous_usage[:responses] - previous_usage[:monthly], 0].max
-    updated_credits = current_credits - consumed_topup_credits - previous_usage[:monthly] + new_plan_credits
+    updated_credits = [current_credits - consumed_topup_credits - previous_usage[:monthly] + new_plan_credits, 0].max
 
-    Rails.logger.info("Updating subscription credits for account #{account.id}: #{current_credits} -> #{updated_credits}")
+    Rails.logger.info("Updating captain credits for account #{account.id}: #{current_credits} -> #{updated_credits}")
     account.update!(limits: current_limits.merge('captain_responses' => updated_credits))
   end
 
diff --git a/spec/enterprise/services/enterprise/billing/create_stripe_customer_service_spec.rb b/spec/enterprise/services/enterprise/billing/create_stripe_customer_service_spec.rb
index 0dd8189c4..d2dbf646a 100644
--- a/spec/enterprise/services/enterprise/billing/create_stripe_customer_service_spec.rb
+++ b/spec/enterprise/services/enterprise/billing/create_stripe_customer_service_spec.rb
@@ -145,10 +145,10 @@ describe Enterprise::Billing::CreateStripeCustomerService do
         account.update!(custom_attributes: { stripe_customer_id: stripe_customer_id })
       end
 
-      context 'when customer has active subscriptions' do
+      context 'when customer has an active non-default subscription' do
         before do
           allow(Stripe::Subscription).to receive(:list).and_return(subscriptions_list)
-          allow(subscriptions_list).to receive(:data).and_return(['subscription'])
+          allow(subscriptions_list).to receive(:data).and_return([{ 'plan' => { 'id' => 'price_paid_plan' } }])
           allow(Stripe::Subscription).to receive(:create)
         end
 

From 5c6ea78ce655a10963c46ca7d077feeb5fd8a4ed Mon Sep 17 00:00:00 2001
From: Muhsin Keloth <muhsinkeramam@gmail.com>
Date: Fri, 8 May 2026 11:42:23 +0400
Subject: [PATCH 04/20] fix(security): Enforce admin authorization on custom
 attribute definitions API (#14392)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Custom attribute definitions can now only be created, edited, or deleted
by administrators, matching the existing settings UI restriction.
Previously, an agent could call the `custom_attribute_definitions` API
directly and modify account configuration that they couldn't reach
through the dashboard — a Broken Access Control vulnerability reported
externally.

Fixes
https://linear.app/chatwoot/issue/CW-7038/broken-access-control-on-custom-attribute-definitions-api

## How to test

1. Sign in as an agent.
2. Try to create a custom attribute by calling `POST
/api/v1/accounts/<id>/custom_attribute_definitions` directly (the
settings page is hidden for agents — use curl with the agent's
`api_access_token`).
3. Expect `401 Unauthorized` with body `{"error":"You are not authorized
to do this action"}`. Repeat for `PATCH` and `DELETE`.
4. Sign in as an administrator and confirm create/edit/delete still work
from Settings → Custom Attributes.
5. As either role, the listing endpoint (`GET
.../custom_attribute_definitions`) should still succeed — agents need
this to render attributes in conversation and contact panels.

Co-authored-by: Muhsin <12408980+muhsin-k@users.noreply.github.com>
---
 ...custom_attribute_definitions_controller.rb |  1 +
 .../custom_attribute_definition_policy.rb     | 21 ++++++++
 ...m_attribute_definitions_controller_spec.rb | 51 ++++++++++++++++---
 3 files changed, 66 insertions(+), 7 deletions(-)
 create mode 100644 app/policies/custom_attribute_definition_policy.rb

diff --git a/app/controllers/api/v1/accounts/custom_attribute_definitions_controller.rb b/app/controllers/api/v1/accounts/custom_attribute_definitions_controller.rb
index 69df99e14..7bd9ae0c3 100644
--- a/app/controllers/api/v1/accounts/custom_attribute_definitions_controller.rb
+++ b/app/controllers/api/v1/accounts/custom_attribute_definitions_controller.rb
@@ -1,6 +1,7 @@
 class Api::V1::Accounts::CustomAttributeDefinitionsController < Api::V1::Accounts::BaseController
   before_action :fetch_custom_attributes_definitions, except: [:create]
   before_action :fetch_custom_attribute_definition, only: [:show, :update, :destroy]
+  before_action :check_authorization
   DEFAULT_ATTRIBUTE_MODEL = 'conversation_attribute'.freeze
 
   def index; end
diff --git a/app/policies/custom_attribute_definition_policy.rb b/app/policies/custom_attribute_definition_policy.rb
new file mode 100644
index 000000000..c62a04b20
--- /dev/null
+++ b/app/policies/custom_attribute_definition_policy.rb
@@ -0,0 +1,21 @@
+class CustomAttributeDefinitionPolicy < ApplicationPolicy
+  def index?
+    @account_user.administrator? || @account_user.agent?
+  end
+
+  def show?
+    @account_user.administrator? || @account_user.agent?
+  end
+
+  def create?
+    @account_user.administrator?
+  end
+
+  def update?
+    @account_user.administrator?
+  end
+
+  def destroy?
+    @account_user.administrator?
+  end
+end
diff --git a/spec/controllers/api/v1/accounts/custom_attribute_definitions_controller_spec.rb b/spec/controllers/api/v1/accounts/custom_attribute_definitions_controller_spec.rb
index c5be392b6..b55da5059 100644
--- a/spec/controllers/api/v1/accounts/custom_attribute_definitions_controller_spec.rb
+++ b/spec/controllers/api/v1/accounts/custom_attribute_definitions_controller_spec.rb
@@ -2,7 +2,8 @@ require 'rails_helper'
 
 RSpec.describe 'Custom Attribute Definitions API', type: :request do
   let(:account) { create(:account) }
-  let(:user) { create(:user, account: account) }
+  let(:agent) { create(:user, account: account, role: :agent) }
+  let(:admin) { create(:user, account: account, role: :administrator) }
 
   describe 'GET /api/v1/accounts/{account.id}/custom_attribute_definitions' do
     context 'when it is an unauthenticated user' do
@@ -19,7 +20,7 @@ RSpec.describe 'Custom Attribute Definitions API', type: :request do
         create(:custom_attribute_definition, attribute_model: 'contact_attribute', account: account)
 
         get "/api/v1/accounts/#{account.id}/custom_attribute_definitions",
-            headers: user.create_new_auth_token,
+            headers: admin.create_new_auth_token,
             as: :json
 
         expect(response).to have_http_status(:success)
@@ -45,7 +46,7 @@ RSpec.describe 'Custom Attribute Definitions API', type: :request do
     context 'when it is an authenticated user' do
       it 'shows the custom attribute definition' do
         get "/api/v1/accounts/#{account.id}/custom_attribute_definitions/#{custom_attribute_definition.id}",
-            headers: user.create_new_auth_token,
+            headers: admin.create_new_auth_token,
             as: :json
 
         expect(response).to have_http_status(:success)
@@ -81,7 +82,7 @@ RSpec.describe 'Custom Attribute Definitions API', type: :request do
     context 'when it is an authenticated user' do
       it 'creates the filter' do
         expect do
-          post "/api/v1/accounts/#{account.id}/custom_attribute_definitions", headers: user.create_new_auth_token,
+          post "/api/v1/accounts/#{account.id}/custom_attribute_definitions", headers: admin.create_new_auth_token,
                                                                               params: payload
         end.to change(CustomAttributeDefinition, :count).by(1)
 
@@ -90,6 +91,18 @@ RSpec.describe 'Custom Attribute Definitions API', type: :request do
         expect(json_response['attribute_key']).to eq 'developer_id'
       end
 
+      context 'when it is an agent' do
+        it 'returns forbidden and does not create the custom attribute' do
+          expect do
+            post "/api/v1/accounts/#{account.id}/custom_attribute_definitions",
+                 headers: agent.create_new_auth_token,
+                 params: payload
+          end.not_to change(CustomAttributeDefinition, :count)
+
+          expect(response).to have_http_status(:unauthorized)
+        end
+      end
+
       context 'when creating with a conflicting attribute_key' do
         let(:standard_key) { CustomAttributeDefinition::STANDARD_ATTRIBUTES[:conversation].first }
         let(:conflicting_payload) do
@@ -105,7 +118,7 @@ RSpec.describe 'Custom Attribute Definitions API', type: :request do
 
         it 'returns error for conflicting key' do
           post "/api/v1/accounts/#{account.id}/custom_attribute_definitions",
-               headers: user.create_new_auth_token,
+               headers: admin.create_new_auth_token,
                params: conflicting_payload
 
           expect(response).to have_http_status(:unprocessable_entity)
@@ -132,7 +145,7 @@ RSpec.describe 'Custom Attribute Definitions API', type: :request do
     context 'when it is an authenticated user' do
       it 'updates the custom attribute definition' do
         patch "/api/v1/accounts/#{account.id}/custom_attribute_definitions/#{custom_attribute_definition.id}",
-              headers: user.create_new_auth_token,
+              headers: admin.create_new_auth_token,
               params: payload,
               as: :json
         expect(response).to have_http_status(:success)
@@ -141,6 +154,19 @@ RSpec.describe 'Custom Attribute Definitions API', type: :request do
         expect(custom_attribute_definition.reload.attribute_model).to eq('conversation_attribute')
       end
     end
+
+    context 'when it is an agent' do
+      it 'returns forbidden and does not update the custom attribute' do
+        original_name = custom_attribute_definition.attribute_display_name
+        patch "/api/v1/accounts/#{account.id}/custom_attribute_definitions/#{custom_attribute_definition.id}",
+              headers: agent.create_new_auth_token,
+              params: payload,
+              as: :json
+
+        expect(response).to have_http_status(:unauthorized)
+        expect(custom_attribute_definition.reload.attribute_display_name).to eq(original_name)
+      end
+    end
   end
 
   describe 'DELETE /api/v1/accounts/{account.id}/custom_attribute_definitions/:id' do
@@ -156,11 +182,22 @@ RSpec.describe 'Custom Attribute Definitions API', type: :request do
     context 'when it is an authenticated admin user' do
       it 'deletes custom attribute' do
         delete "/api/v1/accounts/#{account.id}/custom_attribute_definitions/#{custom_attribute_definition.id}",
-               headers: user.create_new_auth_token,
+               headers: admin.create_new_auth_token,
                as: :json
         expect(response).to have_http_status(:no_content)
         expect(account.custom_attribute_definitions.count).to be 0
       end
     end
+
+    context 'when it is an agent' do
+      it 'returns forbidden and does not delete the custom attribute' do
+        delete "/api/v1/accounts/#{account.id}/custom_attribute_definitions/#{custom_attribute_definition.id}",
+               headers: agent.create_new_auth_token,
+               as: :json
+
+        expect(response).to have_http_status(:unauthorized)
+        expect(account.custom_attribute_definitions.count).to be 1
+      end
+    end
   end
 end

From 9c1d1c4070a26864839798b5718155263f754f1a Mon Sep 17 00:00:00 2001
From: Sojan Jose <sojan@pepalo.com>
Date: Fri, 8 May 2026 13:40:36 +0530
Subject: [PATCH 05/20] feat(labels): remove label associations asynchronously
 on delete (#13531)

## Summary
- Remove label deletion dependency on association cleanup by deleting
immediately and enqueueing a background job.
- Add `Labels::RemoveAssociationsJob` to strip deleted label references
from tagged conversations and contacts.
- Keep this version simple by removing the label count/prompt
requirement requested.

## Implementation notes
- Enqueue job from `Api::V1::Accounts::LabelsController#destroy` with
label title + account id.
- Background work performed in `Labels::DestroyService`.

## References
- Linear issue:
https://linear.app/chatwoot/issue/CW-4765/cw-2857-enhancement-removing-labels-is-inconsistent
- GitHub issue: https://github.com/chatwoot/chatwoot/issues/1249

## Testing
- `bundle exec rspec
spec/controllers/api/v1/accounts/labels_controller_spec.rb
spec/services/labels/destroy_service_spec.rb
spec/jobs/labels/remove_associations_job_spec.rb
spec/services/labels/update_service_spec.rb`
- `bundle exec rubocop
app/controllers/api/v1/accounts/labels_controller.rb
app/jobs/labels/remove_associations_job.rb
spec/controllers/api/v1/accounts/labels_controller_spec.rb
spec/jobs/labels/remove_associations_job_spec.rb
spec/services/labels/destroy_service_spec.rb`

---------

Co-authored-by: Sony Mathew <sony@chatwoot.com>
Co-authored-by: Sony Mathew <2040199+sony-mathew@users.noreply.github.com>
---
 .../api/v1/accounts/labels_controller.rb      |   9 ++
 app/jobs/labels/remove_associations_job.rb    |  11 ++
 app/services/labels/destroy_service.rb        |  60 +++++++++++
 .../api/v1/accounts/labels_controller_spec.rb |  36 +++++++
 .../labels/remove_associations_job_spec.rb    |  21 ++++
 spec/services/labels/destroy_service_spec.rb  | 102 ++++++++++++++++++
 6 files changed, 239 insertions(+)
 create mode 100644 app/jobs/labels/remove_associations_job.rb
 create mode 100644 app/services/labels/destroy_service.rb
 create mode 100644 spec/jobs/labels/remove_associations_job_spec.rb
 create mode 100644 spec/services/labels/destroy_service_spec.rb

diff --git a/app/controllers/api/v1/accounts/labels_controller.rb b/app/controllers/api/v1/accounts/labels_controller.rb
index 54455943b..6889d30a4 100644
--- a/app/controllers/api/v1/accounts/labels_controller.rb
+++ b/app/controllers/api/v1/accounts/labels_controller.rb
@@ -18,7 +18,16 @@ class Api::V1::Accounts::LabelsController < Api::V1::Accounts::BaseController
   end
 
   def destroy
+    label_title = @label.title
+    account_id = Current.account.id
+    label_deleted_at = Time.current
+
     @label.destroy!
+    Labels::RemoveAssociationsJob.perform_later(
+      label_title: label_title,
+      account_id: account_id,
+      label_deleted_at: label_deleted_at
+    )
     head :ok
   end
 
diff --git a/app/jobs/labels/remove_associations_job.rb b/app/jobs/labels/remove_associations_job.rb
new file mode 100644
index 000000000..502d9339e
--- /dev/null
+++ b/app/jobs/labels/remove_associations_job.rb
@@ -0,0 +1,11 @@
+class Labels::RemoveAssociationsJob < ApplicationJob
+  queue_as :default
+
+  def perform(label_title:, account_id:, label_deleted_at:)
+    Labels::DestroyService.new(
+      label_title: label_title,
+      account_id: account_id,
+      label_deleted_at: label_deleted_at
+    ).perform
+  end
+end
diff --git a/app/services/labels/destroy_service.rb b/app/services/labels/destroy_service.rb
new file mode 100644
index 000000000..080e708e7
--- /dev/null
+++ b/app/services/labels/destroy_service.rb
@@ -0,0 +1,60 @@
+class Labels::DestroyService
+  pattr_initialize [:label_title!, :account_id!, :label_deleted_at!]
+
+  def perform
+    remove_conversation_labels
+    remove_contact_labels
+  end
+
+  private
+
+  def remove_conversation_labels
+    tagged_conversations.find_in_batches do |conversation_batch|
+      conversation_batch.each do |conversation|
+        update_conversation_cached_labels(conversation)
+      end
+      delete_label_taggings('Conversation', conversation_batch.map(&:id))
+    end
+  end
+
+  def remove_contact_labels
+    contact_label_taggings.in_batches do |tagging_batch|
+      ActsAsTaggableOn::Tagging.where(id: tagging_batch.select(:id)).delete_all
+    end
+  end
+
+  def update_conversation_cached_labels(conversation)
+    label_list = conversation.label_list.dup
+    label_list.remove(label_title)
+
+    # We only want the acts-as-taggable-on cache effect here, not Conversation callbacks/events.
+    # rubocop:disable Rails/SkipsModelValidations
+    conversation.update_column(:cached_label_list, label_list.join("#{ActsAsTaggableOn.delimiter} "))
+    # rubocop:enable Rails/SkipsModelValidations
+  end
+
+  def tagged_conversations
+    account.conversations.where(id: label_taggings_for('Conversation').select(:taggable_id))
+  end
+
+  def contact_label_taggings
+    label_taggings_for('Contact').where(taggable_id: account.contacts.select(:id))
+  end
+
+  def delete_label_taggings(taggable_type, taggable_ids)
+    ActsAsTaggableOn::Tagging
+      .where(id: label_taggings_for(taggable_type).where(taggable_id: taggable_ids).select(:id))
+      .delete_all
+  end
+
+  def label_taggings_for(taggable_type)
+    ActsAsTaggableOn::Tagging
+      .joins(:tag)
+      .where(context: 'labels', taggable_type: taggable_type, tags: { name: label_title })
+      .where('taggings.created_at <= ?', label_deleted_at)
+  end
+
+  def account
+    @account ||= Account.find(account_id)
+  end
+end
diff --git a/spec/controllers/api/v1/accounts/labels_controller_spec.rb b/spec/controllers/api/v1/accounts/labels_controller_spec.rb
index 61751ad68..657b53796 100644
--- a/spec/controllers/api/v1/accounts/labels_controller_spec.rb
+++ b/spec/controllers/api/v1/accounts/labels_controller_spec.rb
@@ -3,6 +3,7 @@ require 'rails_helper'
 RSpec.describe 'Label API', type: :request do
   let!(:account) { create(:account) }
   let!(:label) { create(:label, account: account) }
+  let!(:conversation) { create(:conversation, account: account) }
 
   describe 'GET /api/v1/accounts/{account.id}/labels' do
     context 'when it is an unauthenticated user' do
@@ -101,4 +102,39 @@ RSpec.describe 'Label API', type: :request do
       end
     end
   end
+
+  describe 'DELETE /api/v1/accounts/{account.id}/labels/:id' do
+    context 'when it is an unauthenticated user' do
+      it 'returns unauthorized' do
+        delete "/api/v1/accounts/#{account.id}/labels/#{label.id}"
+
+        expect(response).to have_http_status(:unauthorized)
+      end
+    end
+
+    context 'when it is an authenticated user' do
+      let(:admin) { create(:user, account: account, role: :administrator) }
+
+      it 'deletes the label and enqueues label cleanup' do
+        label_deleted_at = Time.zone.parse('2026-05-07 10:00:00 UTC')
+        conversation.label_list.add(label.title)
+        conversation.save!
+
+        clear_enqueued_jobs
+
+        travel_to(label_deleted_at) do
+          expect do
+            delete "/api/v1/accounts/#{account.id}/labels/#{label.id}", headers: admin.create_new_auth_token, as: :json
+          end.to have_enqueued_job(Labels::RemoveAssociationsJob).with(
+            label_title: label.title,
+            account_id: account.id,
+            label_deleted_at: label_deleted_at
+          )
+        end
+
+        expect(response).to have_http_status(:ok)
+        expect(Label.exists?(label.id)).to be(false)
+      end
+    end
+  end
 end
diff --git a/spec/jobs/labels/remove_associations_job_spec.rb b/spec/jobs/labels/remove_associations_job_spec.rb
new file mode 100644
index 000000000..0b608630f
--- /dev/null
+++ b/spec/jobs/labels/remove_associations_job_spec.rb
@@ -0,0 +1,21 @@
+require 'rails_helper'
+
+RSpec.describe Labels::RemoveAssociationsJob do
+  subject(:job) do
+    described_class.perform_later(
+      label_title: label_title,
+      account_id: account_id,
+      label_deleted_at: label_deleted_at
+    )
+  end
+
+  let(:label_title) { 'billing' }
+  let(:account_id) { 1 }
+  let(:label_deleted_at) { Time.current }
+
+  it 'queues the job' do
+    expect { job }.to have_enqueued_job(described_class)
+      .with(label_title: label_title, account_id: account_id, label_deleted_at: label_deleted_at)
+      .on_queue('default')
+  end
+end
diff --git a/spec/services/labels/destroy_service_spec.rb b/spec/services/labels/destroy_service_spec.rb
new file mode 100644
index 000000000..7d06b72d0
--- /dev/null
+++ b/spec/services/labels/destroy_service_spec.rb
@@ -0,0 +1,102 @@
+require 'rails_helper'
+
+describe Labels::DestroyService do
+  let(:account) { create(:account) }
+  let(:conversation) { create(:conversation, account: account) }
+  let(:label) { create(:label, account: account) }
+  let(:contact) { conversation.contact }
+  let(:label_deleted_at) { Time.zone.parse('2026-05-07 10:00:00 UTC') }
+
+  before do
+    conversation.label_list.add(label.title)
+    conversation.label_list.add('billing')
+    conversation.save!
+
+    contact.label_list.add(label.title)
+    contact.label_list.add('vip')
+    contact.save!
+
+    set_label_tagging_created_at(conversation, label_deleted_at - 1.minute)
+    set_label_tagging_created_at(contact, label_deleted_at - 1.minute)
+  end
+
+  describe '#perform' do
+    it 'removes label from associated conversations and contacts' do
+      described_class.new(
+        label_title: label.title,
+        account_id: account.id,
+        label_deleted_at: label_deleted_at
+      ).perform
+
+      expect(conversation.reload.label_list).to eq(['billing'])
+      expect(conversation.cached_label_list).to eq('billing')
+      expect(contact.reload.label_list).to eq(['vip'])
+    end
+
+    it 'removes label associations after the label record is destroyed' do
+      label_title = label.title
+      label.destroy!
+
+      described_class.new(
+        label_title: label_title,
+        account_id: account.id,
+        label_deleted_at: label_deleted_at
+      ).perform
+
+      expect(conversation.reload.label_list).to eq(['billing'])
+      expect(conversation.cached_label_list).to eq('billing')
+      expect(contact.reload.label_list).to eq(['vip'])
+    end
+
+    it 'does not remove labels from other accounts' do
+      other_account = create(:account)
+      other_conversation = create(:conversation, account: other_account)
+      other_conversation.label_list.add(label.title)
+      other_conversation.save!
+      set_label_tagging_created_at(other_conversation, label_deleted_at - 1.minute)
+
+      described_class.new(
+        label_title: label.title,
+        account_id: account.id,
+        label_deleted_at: label_deleted_at
+      ).perform
+
+      expect(other_conversation.reload.label_list).to eq([label.title])
+    end
+
+    it 'does not dispatch conversation or contact update events' do
+      expect(Rails.configuration.dispatcher).not_to receive(:dispatch)
+
+      described_class.new(
+        label_title: label.title,
+        account_id: account.id,
+        label_deleted_at: label_deleted_at
+      ).perform
+    end
+
+    it 'does not remove label associations created after the label was deleted' do
+      other_conversation = create(:conversation, account: account)
+      other_conversation.label_list.add(label.title)
+      other_conversation.save!
+      set_label_tagging_created_at(other_conversation, label_deleted_at + 1.minute)
+
+      described_class.new(
+        label_title: label.title,
+        account_id: account.id,
+        label_deleted_at: label_deleted_at
+      ).perform
+
+      expect(conversation.reload.label_list).to eq(['billing'])
+      expect(conversation.cached_label_list).to eq('billing')
+      expect(contact.reload.label_list).to eq(['vip'])
+      expect(other_conversation.reload.label_list).to eq([label.title])
+    end
+  end
+
+  def set_label_tagging_created_at(record, created_at)
+    ActsAsTaggableOn::Tagging
+      .joins(:tag)
+      .find_by!(context: 'labels', taggable: record, tags: { name: label.title })
+      .update!(created_at: created_at)
+  end
+end

From 202403873d24efc4847d212a80314b29f174f2dd Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jo=C3=A3o=20Santos?= <jcps@ua.pt>
Date: Fri, 8 May 2026 12:10:15 +0100
Subject: [PATCH 06/20] feat: Ability to specify the authentication type for
 imap server (#12306)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

# Pull Request Template

## Description

This PR adds IMAP authentication mechanism selection to Chatwoot's email
inbox configuration. Users can now choose between 'plain', 'login', and
'cram-md5' authentication methods when configuring IMAP settings,
providing flexibility for different email providers that require
specific authentication types.

https://github.com/chatwoot/chatwoot/issues/8867

The implementation includes:
- Frontend dropdown with numeric keys (1, 2, 3) matching SMTP auth style
- Backend API validation for allowed authentication mechanisms
- Consistent 'cram-md5' format throughout the codebase
- Updated IMAP service to handle different auth types properly

This feature maintains consistency with existing SMTP authentication
options and follows the established UI/UX patterns in the application.

## Type of change

Please delete options that are not relevant.

- [x] New feature (non-breaking change which adds functionality)
- [ ] Bug fix (non-breaking change which fixes an issue)
- [ ] Breaking change (fix or feature that would cause existing
functionality not to work as expected)
- [ ] This change requires a documentation update

## How Has This Been Tested?

### Manual Testing:
- Tested in Docker environment
- Verified IMAP auth dropdown appears in inbox settings
- Confirmed all three auth mechanisms (plain, login, cram-md5) can be
selected and saved
- Tested API validation by attempting to save invalid auth mechanisms

### Automated Testing:
- Updated existing IMAP service tests to use consistent lowercase values
- Updated API controller tests for authentication parameter handling
- All tests pass locally with the new changes

### Test Configuration:
- Tested with both new and existing inbox configurations

## Checklist:

- [x] My code follows the style guidelines of this project
- [x] I have performed a self-review of my code
- [x] I have commented on my code, particularly in hard-to-understand
areas
- [x] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
- [x] I have added tests that prove my fix is effective or that my
feature works
- [x] New and existing unit tests pass locally with my changes
- [x] Any dependent changes have been merged and published in downstream
modules

## Additional Notes

- This feature is backward compatible and doesn't break existing IMAP
configurations
- The 'cram-md5' format is used consistently throughout (UI, API,
storage, services)
- Net::IMAP compatibility is maintained by converting to 'CRAM-MD5'
internally
- Follows the same pattern established by SMTP authentication
configuration

---------

Co-authored-by: João Santos <joao.santos@madigital.eu>
Co-authored-by: Sony Mathew <sony@chatwoot.com>
---
 app/helpers/api/v1/inboxes_helper.rb          | 28 +++++----
 .../dashboard/i18n/locale/en/inboxMgmt.json   |  3 +-
 .../dashboard/settings/inbox/ImapSettings.vue | 21 +++++++
 app/models/channel/email.rb                   |  3 +-
 app/services/imap/authentication.rb           | 31 ++++++++++
 app/services/imap/base_fetch_email_service.rb |  5 +-
 app/services/imap/fetch_email_service.rb      |  2 +-
 app/views/api/v1/models/_inbox.json.jbuilder  |  1 +
 ...dd_imap_authentication_to_channel_email.rb |  5 ++
 db/schema.rb                                  |  3 +-
 .../v1/accounts/inboxes_controller_spec.rb    | 10 +++-
 spec/factories/channel/channel_email.rb       |  2 +
 .../services/imap/fetch_email_service_spec.rb | 58 ++++++++++++++++++-
 13 files changed, 151 insertions(+), 21 deletions(-)
 create mode 100644 app/services/imap/authentication.rb
 create mode 100644 db/migrate/20260507000000_add_imap_authentication_to_channel_email.rb

diff --git a/app/helpers/api/v1/inboxes_helper.rb b/app/helpers/api/v1/inboxes_helper.rb
index 3d6b559c8..8a10fa99c 100644
--- a/app/helpers/api/v1/inboxes_helper.rb
+++ b/app/helpers/api/v1/inboxes_helper.rb
@@ -17,15 +17,12 @@ module Api::V1::InboxesHelper
   def validate_imap(channel_data)
     return unless channel_data.key?('imap_enabled') && channel_data[:imap_enabled]
 
-    Mail.defaults do
-      retriever_method :imap, { address: channel_data[:imap_address],
-                                port: channel_data[:imap_port],
-                                user_name: channel_data[:imap_login],
-                                password: channel_data[:imap_password],
-                                enable_ssl: channel_data[:imap_enable_ssl] }
-    end
+    # Validate the user-selected auth mechanism before opening the connection.
+    authentication = Imap::Authentication.validate_user_configurable!(channel_data[:imap_authentication])
 
-    check_imap_connection(channel_data)
+    # Use the same auth adapter as the fetch service so LOGIN uses the IMAP LOGIN command,
+    # not SASL AUTH=LOGIN.
+    check_imap_connection(channel_data, authentication)
   end
 
   def validate_smtp(channel_data)
@@ -37,8 +34,8 @@ module Api::V1::InboxesHelper
     check_smtp_connection(channel_data, smtp)
   end
 
-  def check_imap_connection(channel_data)
-    Mail.connection {} # rubocop:disable:block
+  def check_imap_connection(channel_data, authentication)
+    imap = open_imap_connection(channel_data, authentication)
   rescue SocketError => e
     raise StandardError, I18n.t('errors.inboxes.imap.socket_error')
   rescue Net::IMAP::NoResponseError => e
@@ -53,9 +50,20 @@ module Api::V1::InboxesHelper
   rescue StandardError => e
     raise StandardError, e.message
   ensure
+    imap.disconnect if imap.present? && !imap.disconnected?
     Rails.logger.error "[Api::V1::InboxesHelper] check_imap_connection failed with #{e.message}" if e.present?
   end
 
+  def open_imap_connection(channel_data, authentication)
+    imap = build_imap_connection(channel_data)
+    Imap::Authentication.authenticate!(imap, authentication, channel_data[:imap_login], channel_data[:imap_password])
+    imap
+  end
+
+  def build_imap_connection(channel_data)
+    Net::IMAP.new(channel_data[:imap_address], port: channel_data[:imap_port], ssl: channel_data[:imap_enable_ssl])
+  end
+
   def check_smtp_connection(channel_data, smtp)
     smtp.open_timeout = 10
     smtp.start(channel_data[:smtp_domain], channel_data[:smtp_login], channel_data[:smtp_password],
diff --git a/app/javascript/dashboard/i18n/locale/en/inboxMgmt.json b/app/javascript/dashboard/i18n/locale/en/inboxMgmt.json
index 24207cef2..2620f89f7 100644
--- a/app/javascript/dashboard/i18n/locale/en/inboxMgmt.json
+++ b/app/javascript/dashboard/i18n/locale/en/inboxMgmt.json
@@ -1011,7 +1011,8 @@
         "LABEL": "Password",
         "PLACE_HOLDER": "Password"
       },
-      "ENABLE_SSL": "Enable SSL"
+      "ENABLE_SSL": "Enable SSL",
+      "AUTH_MECHANISM": "Authentication"
     },
     "MICROSOFT": {
       "TITLE": "Microsoft",
diff --git a/app/javascript/dashboard/routes/dashboard/settings/inbox/ImapSettings.vue b/app/javascript/dashboard/routes/dashboard/settings/inbox/ImapSettings.vue
index 7325793db..ab25ed1af 100644
--- a/app/javascript/dashboard/routes/dashboard/settings/inbox/ImapSettings.vue
+++ b/app/javascript/dashboard/routes/dashboard/settings/inbox/ImapSettings.vue
@@ -5,11 +5,13 @@ import SettingsFieldSection from 'dashboard/components-next/Settings/SettingsFie
 import { useVuelidate } from '@vuelidate/core';
 import { required, minLength } from '@vuelidate/validators';
 import NextButton from 'dashboard/components-next/button/Button.vue';
+import SingleSelectDropdown from './components/SingleSelectDropdown.vue';
 
 export default {
   components: {
     SettingsFieldSection,
     NextButton,
+    SingleSelectDropdown,
   },
   props: {
     inbox: {
@@ -28,6 +30,12 @@ export default {
       login: '',
       password: '',
       isSSLEnabled: true,
+      authMechanism: 'plain',
+      authMechanisms: [
+        { key: 1, value: 'plain' },
+        { key: 2, value: 'login' },
+        { key: 3, value: 'cram-md5' },
+      ],
     };
   },
   validations: {
@@ -56,6 +64,7 @@ export default {
         imap_login,
         imap_password,
         imap_enable_ssl,
+        imap_authentication,
       } = this.inbox;
       this.isIMAPEnabled = imap_enabled;
       this.address = imap_address;
@@ -63,6 +72,7 @@ export default {
       this.login = imap_login;
       this.password = imap_password;
       this.isSSLEnabled = imap_enable_ssl;
+      this.authMechanism = imap_authentication || 'plain';
     },
     async updateInbox() {
       try {
@@ -77,6 +87,7 @@ export default {
             imap_login: this.login,
             imap_password: this.password,
             imap_enable_ssl: this.isSSLEnabled,
+            imap_authentication: this.authMechanism,
           },
         };
 
@@ -90,6 +101,9 @@ export default {
         useAlert(error.message);
       }
     },
+    handleAuthMechanismChange(mode) {
+      this.authMechanism = mode;
+    },
   },
 };
 </script>
@@ -155,6 +169,13 @@ export default {
           />
           {{ $t('INBOX_MGMT.IMAP.ENABLE_SSL') }}
         </label>
+        <SingleSelectDropdown
+          class="w-full"
+          :label="$t('INBOX_MGMT.IMAP.AUTH_MECHANISM')"
+          :selected="authMechanism"
+          :options="authMechanisms"
+          :action="handleAuthMechanismChange"
+        />
       </div>
       <NextButton
         type="submit"
diff --git a/app/models/channel/email.rb b/app/models/channel/email.rb
index b1124dd75..e61123b22 100644
--- a/app/models/channel/email.rb
+++ b/app/models/channel/email.rb
@@ -6,6 +6,7 @@
 #  email                     :string           not null
 #  forward_to_email          :string           not null
 #  imap_address              :string           default("")
+#  imap_authentication       :string           default("plain")
 #  imap_enable_ssl           :boolean          default(TRUE)
 #  imap_enabled              :boolean          default(FALSE)
 #  imap_login                :string           default("")
@@ -47,7 +48,7 @@ class Channel::Email < ApplicationRecord
   end
 
   self.table_name = 'channel_email'
-  EDITABLE_ATTRS = [:email, :imap_enabled, :imap_login, :imap_password, :imap_address, :imap_port, :imap_enable_ssl,
+  EDITABLE_ATTRS = [:email, :imap_enabled, :imap_login, :imap_password, :imap_address, :imap_port, :imap_enable_ssl, :imap_authentication,
                     :smtp_enabled, :smtp_login, :smtp_password, :smtp_address, :smtp_port, :smtp_domain, :smtp_enable_starttls_auto,
                     :smtp_enable_ssl_tls, :smtp_openssl_verify_mode, :smtp_authentication, :provider, :verified_for_sending].freeze
 
diff --git a/app/services/imap/authentication.rb b/app/services/imap/authentication.rb
new file mode 100644
index 000000000..d5eb17cdd
--- /dev/null
+++ b/app/services/imap/authentication.rb
@@ -0,0 +1,31 @@
+module Imap::Authentication
+  DEFAULT_MECHANISM = 'plain'.freeze
+  USER_CONFIGURABLE_MECHANISMS = %w[plain login cram-md5].freeze
+
+  module_function
+
+  def normalize(mechanism)
+    mechanism.presence || DEFAULT_MECHANISM
+  end
+
+  def validate_user_configurable!(mechanism)
+    normalized_mechanism = normalize(mechanism).to_s.downcase
+    return normalized_mechanism if USER_CONFIGURABLE_MECHANISMS.include?(normalized_mechanism)
+
+    allowed_values = USER_CONFIGURABLE_MECHANISMS.join(', ')
+    raise StandardError, "Invalid IMAP authentication mechanism. Allowed values: #{allowed_values}"
+  end
+
+  def authenticate!(imap, mechanism, username, password)
+    normalized_mechanism = normalize(mechanism).to_s.downcase
+
+    case normalized_mechanism
+    when 'cram-md5'
+      imap.authenticate('CRAM-MD5', username, password)
+    when 'login'
+      imap.login(username, password)
+    else
+      imap.authenticate(normalize(mechanism), username, password)
+    end
+  end
+end
diff --git a/app/services/imap/base_fetch_email_service.rb b/app/services/imap/base_fetch_email_service.rb
index 17114f516..5a6d3537c 100644
--- a/app/services/imap/base_fetch_email_service.rb
+++ b/app/services/imap/base_fetch_email_service.rb
@@ -130,8 +130,9 @@ class Imap::BaseFetchEmailService
   end
 
   def build_imap_client
-    imap = Net::IMAP.new(channel.imap_address, port: channel.imap_port, ssl: true)
-    imap.authenticate(authentication_type, channel.imap_login, imap_password)
+    imap = Net::IMAP.new(channel.imap_address, port: channel.imap_port, ssl: channel.imap_enable_ssl)
+    Imap::Authentication.authenticate!(imap, authentication_type, channel.imap_login, imap_password)
+
     imap.select('INBOX')
     imap
   end
diff --git a/app/services/imap/fetch_email_service.rb b/app/services/imap/fetch_email_service.rb
index f602b3375..2fa389e31 100644
--- a/app/services/imap/fetch_email_service.rb
+++ b/app/services/imap/fetch_email_service.rb
@@ -6,7 +6,7 @@ class Imap::FetchEmailService < Imap::BaseFetchEmailService
   private
 
   def authentication_type
-    'PLAIN'
+    channel.imap_authentication || 'plain'
   end
 
   def imap_password
diff --git a/app/views/api/v1/models/_inbox.json.jbuilder b/app/views/api/v1/models/_inbox.json.jbuilder
index e4c19fea0..3da71d07d 100644
--- a/app/views/api/v1/models/_inbox.json.jbuilder
+++ b/app/views/api/v1/models/_inbox.json.jbuilder
@@ -90,6 +90,7 @@ if resource.email?
     json.imap_port resource.channel.try(:imap_port)
     json.imap_enabled resource.channel.try(:imap_enabled)
     json.imap_enable_ssl resource.channel.try(:imap_enable_ssl)
+    json.imap_authentication resource.channel.try(:imap_authentication)
 
     if resource.channel.try(:microsoft?) || resource.channel.try(:google?) || resource.channel.try(:legacy_google?)
       json.reauthorization_required resource.channel.try(:provider_config).empty? || resource.channel.try(:reauthorization_required?)
diff --git a/db/migrate/20260507000000_add_imap_authentication_to_channel_email.rb b/db/migrate/20260507000000_add_imap_authentication_to_channel_email.rb
new file mode 100644
index 000000000..ffdf94bf2
--- /dev/null
+++ b/db/migrate/20260507000000_add_imap_authentication_to_channel_email.rb
@@ -0,0 +1,5 @@
+class AddImapAuthenticationToChannelEmail < ActiveRecord::Migration[7.0]
+  def change
+    add_column :channel_email, :imap_authentication, :string, default: 'plain'
+  end
+end
diff --git a/db/schema.rb b/db/schema.rb
index f7c3cce6d..9f8d76b78 100644
--- a/db/schema.rb
+++ b/db/schema.rb
@@ -10,7 +10,7 @@
 #
 # It's strongly recommended that you check this file into your version control system.
 
-ActiveRecord::Schema[7.1].define(version: 2026_04_30_114500) do
+ActiveRecord::Schema[7.1].define(version: 2026_05_07_000000) do
   # These extensions should be enabled to support this database
   enable_extension "pg_stat_statements"
   enable_extension "pg_trgm"
@@ -472,6 +472,7 @@ ActiveRecord::Schema[7.1].define(version: 2026_04_30_114500) do
     t.boolean "smtp_enable_ssl_tls", default: false
     t.jsonb "provider_config", default: {}
     t.string "provider"
+    t.string "imap_authentication", default: "plain"
     t.boolean "verified_for_sending", default: false, null: false
     t.index ["email"], name: "index_channel_email_on_email", unique: true
     t.index ["forward_to_email"], name: "index_channel_email_on_forward_to_email", unique: true
diff --git a/spec/controllers/api/v1/accounts/inboxes_controller_spec.rb b/spec/controllers/api/v1/accounts/inboxes_controller_spec.rb
index 8aae60d53..0372bb141 100644
--- a/spec/controllers/api/v1/accounts/inboxes_controller_spec.rb
+++ b/spec/controllers/api/v1/accounts/inboxes_controller_spec.rb
@@ -568,8 +568,10 @@ RSpec.describe 'Inboxes API', type: :request do
         email_channel = create(:channel_email, account: account)
         email_inbox = create(:inbox, channel: email_channel, account: account)
 
-        imap_connection = double
-        allow(Mail).to receive(:connection).and_return(imap_connection)
+        imap_connection = instance_double(Net::IMAP, disconnected?: false)
+        allow(Net::IMAP).to receive(:new).and_return(imap_connection)
+        allow(imap_connection).to receive(:login)
+        allow(imap_connection).to receive(:disconnect)
 
         patch "/api/v1/accounts/#{account.id}/inboxes/#{email_inbox.id}",
               headers: admin.create_new_auth_token,
@@ -578,7 +580,8 @@ RSpec.describe 'Inboxes API', type: :request do
                   imap_enabled: true,
                   imap_address: 'imap.gmail.com',
                   imap_port: 993,
-                  imap_login: 'imaptest@gmail.com'
+                  imap_login: 'imaptest@gmail.com',
+                  imap_authentication: 'login'
                 }
               },
               as: :json
@@ -587,6 +590,7 @@ RSpec.describe 'Inboxes API', type: :request do
         expect(email_channel.reload.imap_enabled).to be true
         expect(email_channel.reload.imap_address).to eq('imap.gmail.com')
         expect(email_channel.reload.imap_port).to eq(993)
+        expect(email_channel.reload.imap_authentication).to eq('login')
       end
 
       it 'updates avatar when administrator' do
diff --git a/spec/factories/channel/channel_email.rb b/spec/factories/channel/channel_email.rb
index 32cdb0488..516699594 100644
--- a/spec/factories/channel/channel_email.rb
+++ b/spec/factories/channel/channel_email.rb
@@ -16,6 +16,7 @@ FactoryBot.define do
       imap_login { 'email@example.com' }
       imap_password { '' }
       imap_enable_ssl { true }
+
       provider_config do
         {
           expires_on: Time.zone.now + 3600,
@@ -33,6 +34,7 @@ FactoryBot.define do
       imap_login { 'email@example.com' }
       imap_password { 'random-password' }
       imap_enable_ssl { true }
+      imap_authentication { 'plain' }
     end
   end
 end
diff --git a/spec/services/imap/fetch_email_service_spec.rb b/spec/services/imap/fetch_email_service_spec.rb
index 2910f41fb..a40a46343 100644
--- a/spec/services/imap/fetch_email_service_spec.rb
+++ b/spec/services/imap/fetch_email_service_spec.rb
@@ -13,14 +13,68 @@ RSpec.describe Imap::FetchEmailService do
     before do
       allow(Rails).to receive(:logger).and_return(logger)
       allow(Net::IMAP).to receive(:new).with(
-        imap_email_channel.imap_address, port: imap_email_channel.imap_port, ssl: true
+        imap_email_channel.imap_address, port: imap_email_channel.imap_port, ssl: imap_email_channel.imap_enable_ssl
       ).and_return(imap)
       allow(imap).to receive(:authenticate).with(
-        'PLAIN', imap_email_channel.imap_login, imap_email_channel.imap_password
+        'plain', imap_email_channel.imap_login, imap_email_channel.imap_password
       )
       allow(imap).to receive(:select).with('INBOX')
     end
 
+    context 'when using CRAM-MD5 authentication' do
+      let(:cram_md5_channel) { create(:channel_email, :imap_email, account: account, imap_authentication: 'cram-md5') }
+
+      before do
+        allow(Net::IMAP).to receive(:new).with(
+          cram_md5_channel.imap_address, port: cram_md5_channel.imap_port, ssl: cram_md5_channel.imap_enable_ssl
+        ).and_return(imap)
+        allow(imap).to receive(:authenticate).with(
+          'CRAM-MD5', cram_md5_channel.imap_login, cram_md5_channel.imap_password
+        )
+        allow(imap).to receive(:select).with('INBOX')
+      end
+
+      it 'uses CRAM-MD5 authentication' do
+        travel_to '26.10.2020 10:00'.to_datetime do
+          allow(imap).to receive(:search).with(%w[SINCE 25-Oct-2020]).and_return([])
+          allow(imap).to receive(:logout)
+
+          described_class.new(channel: cram_md5_channel).perform
+
+          expect(imap).to have_received(:authenticate).with(
+            'CRAM-MD5', cram_md5_channel.imap_login, cram_md5_channel.imap_password
+          )
+        end
+      end
+    end
+
+    context 'when using LOGIN authentication' do
+      let(:login_channel) { create(:channel_email, :imap_email, account: account, imap_authentication: 'login') }
+
+      before do
+        allow(Net::IMAP).to receive(:new).with(
+          login_channel.imap_address, port: login_channel.imap_port, ssl: login_channel.imap_enable_ssl
+        ).and_return(imap)
+        allow(imap).to receive(:login).with(
+          login_channel.imap_login, login_channel.imap_password
+        )
+        allow(imap).to receive(:select).with('INBOX')
+      end
+
+      it 'uses LOGIN authentication' do
+        travel_to '26.10.2020 10:00'.to_datetime do
+          allow(imap).to receive(:search).with(%w[SINCE 25-Oct-2020]).and_return([])
+          allow(imap).to receive(:logout)
+
+          described_class.new(channel: login_channel).perform
+
+          expect(imap).to have_received(:login).with(
+            login_channel.imap_login, login_channel.imap_password
+          )
+        end
+      end
+    end
+
     context 'when new emails are available in the mailbox' do
       it 'fetches the emails and returns the emails that are not present in the db' do
         travel_to '26.10.2020 10:00'.to_datetime do

From aa10d42237c9c0d8c2c903144a293bb93b16bdf0 Mon Sep 17 00:00:00 2001
From: Aakash Bakhle <48802744+aakashb95@users.noreply.github.com>
Date: Fri, 8 May 2026 18:44:15 +0530
Subject: [PATCH 07/20] chore: bump RubyLLM version [AI-152] (#14387)

# Pull Request Template

## Description

Bump RubyLLM version and update model registry

## Type of change

Version bump on package

## How Has This Been Tested?

Please describe the tests that you ran to verify your changes. Provide
instructions so we can reproduce. Please also list any relevant details
for your test configuration.

locally and specs


## Checklist:

- [x] My code follows the style guidelines of this project
- [x] I have performed a self-review of my code
- [x] I have commented on my code, particularly in hard-to-understand
areas
- [ ] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
- [x] I have added tests that prove my fix is effective or that my
feature works
- [x] New and existing unit tests pass locally with my changes
- [x] Any dependent changes have been merged and published in downstream
modules
---
 Gemfile                |     4 +-
 Gemfile.lock           |    24 +-
 config/llm_models.json | 42328 ++++++++++++++++++++++++++++++++++++---
 3 files changed, 39802 insertions(+), 2554 deletions(-)

diff --git a/Gemfile b/Gemfile
index 8c0bb82a5..dfcdb3e30 100644
--- a/Gemfile
+++ b/Gemfile
@@ -195,10 +195,10 @@ gem 'reverse_markdown'
 
 gem 'iso-639'
 gem 'ruby-openai'
-gem 'ai-agents', '>= 0.9.1'
+gem 'ai-agents', '>= 0.10.0'
 
 # TODO: Move this gem as a dependency of ai-agents
-gem 'ruby_llm', '>= 1.8.2'
+gem 'ruby_llm', '>= 1.14.1'
 gem 'ruby_llm-schema'
 
 gem 'cld3', '~> 3.7'
diff --git a/Gemfile.lock b/Gemfile.lock
index bd21b7a36..15d9b2614 100644
--- a/Gemfile.lock
+++ b/Gemfile.lock
@@ -126,8 +126,8 @@ GEM
       jbuilder (~> 2)
       rails (>= 4.2, < 7.2)
       selectize-rails (~> 0.6)
-    ai-agents (0.9.1)
-      ruby_llm (~> 1.9.1)
+    ai-agents (0.10.0)
+      ruby_llm (~> 1.14)
     annotaterb (4.20.0)
       activerecord (>= 6.0.0)
       activesupport (>= 6.0.0)
@@ -307,8 +307,8 @@ GEM
     faraday-mashify (1.0.0)
       faraday (~> 2.0)
       hashie
-    faraday-multipart (1.0.4)
-      multipart-post (~> 2)
+    faraday-multipart (1.2.0)
+      multipart-post (~> 2.0)
     faraday-net_http (3.4.2)
       net-http (~> 0.5)
     faraday-net_http_persistent (2.1.0)
@@ -466,7 +466,7 @@ GEM
       rails-dom-testing (>= 1, < 3)
       railties (>= 4.2.0)
       thor (>= 0.14, < 2.0)
-    json (2.19.2)
+    json (2.19.5)
     json_refs (0.1.8)
       hana
     json_schemer (0.2.24)
@@ -835,17 +835,17 @@ GEM
     ruby2ruby (2.5.0)
       ruby_parser (~> 3.1)
       sexp_processor (~> 4.6)
-    ruby_llm (1.9.2)
+    ruby_llm (1.15.0)
       base64
       event_stream_parser (~> 1)
       faraday (>= 1.10.0)
       faraday-multipart (>= 1)
       faraday-net_http (>= 1)
       faraday-retry (>= 1)
-      marcel (~> 1.0)
-      ruby_llm-schema (~> 0.2.1)
+      marcel (~> 1)
+      ruby_llm-schema (~> 0)
       zeitwerk (~> 2)
-    ruby_llm-schema (0.2.5)
+    ruby_llm-schema (0.3.0)
     ruby_parser (3.20.0)
       sexp_processor (~> 4.16)
     sass (3.7.4)
@@ -1019,7 +1019,7 @@ GEM
     working_hours (1.4.1)
       activesupport (>= 3.2)
       tzinfo
-    zeitwerk (2.7.4)
+    zeitwerk (2.7.5)
 
 PLATFORMS
   arm64-darwin-20
@@ -1039,7 +1039,7 @@ DEPENDENCIES
   administrate (>= 0.20.1)
   administrate-field-active_storage (>= 1.0.3)
   administrate-field-belongs_to_search (>= 0.9.0)
-  ai-agents (>= 0.9.1)
+  ai-agents (>= 0.10.0)
   annotaterb
   attr_extras
   audited (~> 5.4, >= 5.4.1)
@@ -1144,7 +1144,7 @@ DEPENDENCIES
   rubocop-rails
   rubocop-rspec
   ruby-openai
-  ruby_llm (>= 1.8.2)
+  ruby_llm (>= 1.14.1)
   ruby_llm-schema
   scout_apm
   scss_lint
diff --git a/config/llm_models.json b/config/llm_models.json
index e2fe0e938..00c2038e5 100644
--- a/config/llm_models.json
+++ b/config/llm_models.json
@@ -4,7 +4,7 @@
     "name": "Claude Haiku 3.5",
     "provider": "anthropic",
     "family": "claude-haiku",
-    "created_at": "2024-10-22 00:00:00 +0530",
+    "created_at": "2024-10-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": "2024-07-31",
@@ -27,7 +27,8 @@
         "standard": {
           "input_per_million": 0.8,
           "output_per_million": 4,
-          "cached_input_per_million": 0.08
+          "cache_read_input_per_million": 0.08,
+          "cache_write_input_per_million": 1
         }
       }
     },
@@ -56,7 +57,7 @@
     "name": "Claude Haiku 3.5 (latest)",
     "provider": "anthropic",
     "family": "claude-haiku",
-    "created_at": "2024-10-22 00:00:00 +0530",
+    "created_at": "2024-10-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": "2024-07-31",
@@ -79,7 +80,8 @@
         "standard": {
           "input_per_million": 0.8,
           "output_per_million": 4,
-          "cached_input_per_million": 0.08
+          "cache_read_input_per_million": 0.08,
+          "cache_write_input_per_million": 1
         }
       }
     },
@@ -108,7 +110,7 @@
     "name": "Claude Sonnet 3.5",
     "provider": "anthropic",
     "family": "claude-sonnet",
-    "created_at": "2024-06-20 00:00:00 +0530",
+    "created_at": "2024-06-20 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": "2024-04-30",
@@ -131,7 +133,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -160,7 +163,7 @@
     "name": "Claude Sonnet 3.5 v2",
     "provider": "anthropic",
     "family": "claude-sonnet",
-    "created_at": "2024-10-22 00:00:00 +0530",
+    "created_at": "2024-10-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": "2024-04-30",
@@ -183,7 +186,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -212,7 +216,7 @@
     "name": "Claude Sonnet 3.7",
     "provider": "anthropic",
     "family": "claude-sonnet",
-    "created_at": "2025-02-19 00:00:00 +0530",
+    "created_at": "2025-02-19 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2024-10-31",
@@ -236,7 +240,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -265,7 +270,7 @@
     "name": "Claude Haiku 3",
     "provider": "anthropic",
     "family": "claude-haiku",
-    "created_at": "2024-03-13 00:00:00 +0530",
+    "created_at": "2024-03-13 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": "2023-08-31",
@@ -288,7 +293,8 @@
         "standard": {
           "input_per_million": 0.25,
           "output_per_million": 1.25,
-          "cached_input_per_million": 0.03
+          "cache_read_input_per_million": 0.03,
+          "cache_write_input_per_million": 0.3
         }
       }
     },
@@ -317,7 +323,7 @@
     "name": "Claude Opus 3",
     "provider": "anthropic",
     "family": "claude-opus",
-    "created_at": "2024-02-29 00:00:00 +0530",
+    "created_at": "2024-02-29 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": "2023-08-31",
@@ -340,7 +346,8 @@
         "standard": {
           "input_per_million": 15,
           "output_per_million": 75,
-          "cached_input_per_million": 1.5
+          "cache_read_input_per_million": 1.5,
+          "cache_write_input_per_million": 18.75
         }
       }
     },
@@ -369,7 +376,7 @@
     "name": "Claude Sonnet 3",
     "provider": "anthropic",
     "family": "claude-sonnet",
-    "created_at": "2024-03-04 00:00:00 +0530",
+    "created_at": "2024-03-04 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": "2023-08-31",
@@ -392,7 +399,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 0.3
         }
       }
     },
@@ -421,7 +429,7 @@
     "name": "Claude Haiku 4.5 (latest)",
     "provider": "anthropic",
     "family": "claude-haiku",
-    "created_at": "2025-10-15 00:00:00 +0530",
+    "created_at": "2025-10-15 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-02-28",
@@ -445,7 +453,8 @@
         "standard": {
           "input_per_million": 1,
           "output_per_million": 5,
-          "cached_input_per_million": 0.1
+          "cache_read_input_per_million": 0.1,
+          "cache_write_input_per_million": 1.25
         }
       }
     },
@@ -474,7 +483,7 @@
     "name": "Claude Haiku 4.5",
     "provider": "anthropic",
     "family": "claude-haiku",
-    "created_at": "2025-10-15 00:00:00 +0530",
+    "created_at": "2025-10-15 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-02-28",
@@ -498,7 +507,8 @@
         "standard": {
           "input_per_million": 1,
           "output_per_million": 5,
-          "cached_input_per_million": 0.1
+          "cache_read_input_per_million": 0.1,
+          "cache_write_input_per_million": 1.25
         }
       }
     },
@@ -527,7 +537,7 @@
     "name": "Claude Opus 4 (latest)",
     "provider": "anthropic",
     "family": "claude-opus",
-    "created_at": "2025-05-22 00:00:00 +0530",
+    "created_at": "2025-05-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 32000,
     "knowledge_cutoff": "2025-03-31",
@@ -551,7 +561,8 @@
         "standard": {
           "input_per_million": 15,
           "output_per_million": 75,
-          "cached_input_per_million": 1.5
+          "cache_read_input_per_million": 1.5,
+          "cache_write_input_per_million": 18.75
         }
       }
     },
@@ -580,7 +591,7 @@
     "name": "Claude Opus 4.1 (latest)",
     "provider": "anthropic",
     "family": "claude-opus",
-    "created_at": "2025-08-05 00:00:00 +0530",
+    "created_at": "2025-08-05 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 32000,
     "knowledge_cutoff": "2025-03-31",
@@ -604,7 +615,8 @@
         "standard": {
           "input_per_million": 15,
           "output_per_million": 75,
-          "cached_input_per_million": 1.5
+          "cache_read_input_per_million": 1.5,
+          "cache_write_input_per_million": 18.75
         }
       }
     },
@@ -633,7 +645,7 @@
     "name": "Claude Opus 4.1",
     "provider": "anthropic",
     "family": "claude-opus",
-    "created_at": "2025-08-05 00:00:00 +0530",
+    "created_at": "2025-08-05 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 32000,
     "knowledge_cutoff": "2025-03-31",
@@ -657,7 +669,8 @@
         "standard": {
           "input_per_million": 15,
           "output_per_million": 75,
-          "cached_input_per_million": 1.5
+          "cache_read_input_per_million": 1.5,
+          "cache_write_input_per_million": 18.75
         }
       }
     },
@@ -686,7 +699,7 @@
     "name": "Claude Opus 4",
     "provider": "anthropic",
     "family": "claude-opus",
-    "created_at": "2025-05-22 00:00:00 +0530",
+    "created_at": "2025-05-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 32000,
     "knowledge_cutoff": "2025-03-31",
@@ -710,7 +723,8 @@
         "standard": {
           "input_per_million": 15,
           "output_per_million": 75,
-          "cached_input_per_million": 1.5
+          "cache_read_input_per_million": 1.5,
+          "cache_write_input_per_million": 18.75
         }
       }
     },
@@ -739,7 +753,7 @@
     "name": "Claude Opus 4.5 (latest)",
     "provider": "anthropic",
     "family": "claude-opus",
-    "created_at": "2025-11-24 00:00:00 +0530",
+    "created_at": "2025-11-24 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-03-31",
@@ -763,7 +777,8 @@
         "standard": {
           "input_per_million": 5,
           "output_per_million": 25,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
         }
       }
     },
@@ -792,7 +807,7 @@
     "name": "Claude Opus 4.5",
     "provider": "anthropic",
     "family": "claude-opus",
-    "created_at": "2025-11-01 00:00:00 +0530",
+    "created_at": "2025-11-01 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-03-31",
@@ -816,7 +831,8 @@
         "standard": {
           "input_per_million": 5,
           "output_per_million": 25,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
         }
       }
     },
@@ -845,10 +861,10 @@
     "name": "Claude Opus 4.6",
     "provider": "anthropic",
     "family": "claude-opus",
-    "created_at": "2026-02-05 00:00:00 +0530",
+    "created_at": "2026-02-05 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 128000,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2025-05-31",
     "modalities": {
       "input": [
         "text",
@@ -869,7 +885,8 @@
         "standard": {
           "input_per_million": 5,
           "output_per_million": 25,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
         }
       }
     },
@@ -890,7 +907,61 @@
         "context": 1000000,
         "output": 128000
       },
-      "knowledge": "2025-05"
+      "knowledge": "2025-05-31"
+    }
+  },
+  {
+    "id": "claude-opus-4-7",
+    "name": "Claude Opus 4.7",
+    "provider": "anthropic",
+    "family": "claude-opus",
+    "created_at": "2026-04-16 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": "2026-01-31",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5,
+          "output_per_million": 25,
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "anthropic",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": false,
+      "last_updated": "2026-04-16",
+      "cost": {
+        "input": 5,
+        "output": 25,
+        "cache_read": 0.5,
+        "cache_write": 6.25
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 128000
+      },
+      "knowledge": "2026-01-31"
     }
   },
   {
@@ -898,7 +969,7 @@
     "name": "Claude Sonnet 4 (latest)",
     "provider": "anthropic",
     "family": "claude-sonnet",
-    "created_at": "2025-05-22 00:00:00 +0530",
+    "created_at": "2025-05-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-03-31",
@@ -922,7 +993,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -951,7 +1023,7 @@
     "name": "Claude Sonnet 4",
     "provider": "anthropic",
     "family": "claude-sonnet",
-    "created_at": "2025-05-22 00:00:00 +0530",
+    "created_at": "2025-05-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-03-31",
@@ -975,7 +1047,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -1004,7 +1077,7 @@
     "name": "Claude Sonnet 4.5 (latest)",
     "provider": "anthropic",
     "family": "claude-sonnet",
-    "created_at": "2025-09-29 00:00:00 +0530",
+    "created_at": "2025-09-29 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-07-31",
@@ -1028,7 +1101,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -1057,7 +1131,7 @@
     "name": "Claude Sonnet 4.5",
     "provider": "anthropic",
     "family": "claude-sonnet",
-    "created_at": "2025-09-29 00:00:00 +0530",
+    "created_at": "2025-09-29 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-07-31",
@@ -1081,7 +1155,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -1110,10 +1185,10 @@
     "name": "Claude Sonnet 4.6",
     "provider": "anthropic",
     "family": "claude-sonnet",
-    "created_at": "2026-02-17 00:00:00 +0530",
+    "created_at": "2026-02-17 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 64000,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2025-08-31",
     "modalities": {
       "input": [
         "text",
@@ -1134,7 +1209,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -1155,7 +1231,8577 @@
         "context": 1000000,
         "output": 64000
       },
-      "knowledge": "2025-08"
+      "knowledge": "2025-08-31"
+    }
+  },
+  {
+    "id": "AI21-Jamba-1.5-Large",
+    "name": "AI21-Jamba-1.5-Large",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "AI21-Jamba-1.5-Mini",
+    "name": "AI21-Jamba-1.5-Mini",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "AI21-Jamba-Instruct",
+    "name": "AI21-Jamba-Instruct",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Codestral-2501-2",
+    "name": "Codestral-2501-2",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Cohere-command-r",
+    "name": "Cohere-command-r",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Cohere-command-r-08-2024",
+    "name": "Cohere-command-r-08-2024",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Cohere-command-r-plus",
+    "name": "Cohere-command-r-plus",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Cohere-command-r-plus-08-2024",
+    "name": "Cohere-command-r-plus-08-2024",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Cohere-embed-v3-english",
+    "name": "Cohere-embed-v3-english",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Cohere-embed-v3-multilingual",
+    "name": "Cohere-embed-v3-multilingual",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Cohere-rerank-v4.0-fast",
+    "name": "Cohere-rerank-v4.0-fast",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Cohere-rerank-v4.0-pro",
+    "name": "Cohere-rerank-v4.0-pro",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "DeepSeek-R1",
+    "name": "DeepSeek-R1",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "DeepSeek-R1-0528",
+    "name": "DeepSeek-R1-0528",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "DeepSeek-V3",
+    "name": "DeepSeek-V3",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "DeepSeek-V3-0324",
+    "name": "DeepSeek-V3-0324",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "DeepSeek-V3.1",
+    "name": "DeepSeek-V3.1",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "DeepSeek-V3.2",
+    "name": "DeepSeek-V3.2",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "DeepSeek-V3.2-Speciale",
+    "name": "DeepSeek-V3.2-Speciale",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "DeepSeek-V4-Flash-2026-04-23",
+    "name": "DeepSeek-V4-Flash-2026-04-23",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "FLUX-1.1-pro",
+    "name": "FLUX-1.1-pro",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "FLUX.1-Kontext-pro",
+    "name": "FLUX.1-Kontext-pro",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "FLUX.2-pro",
+    "name": "FLUX.2-pro",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Kimi-K2-Thinking",
+    "name": "Kimi-K2-Thinking",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Kimi-K2.5",
+    "name": "Kimi-K2.5",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Kimi-K2.6-2026-04-20",
+    "name": "Kimi-K2.6-2026-04-20",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Llama-3.2-11B-Vision-Instruct",
+    "name": "Llama-3.2-11B-Vision-Instruct",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Llama-3.2-11B-Vision-Instruct-2",
+    "name": "Llama-3.2-11B-Vision-Instruct-2",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Llama-3.2-90B-Vision-Instruct",
+    "name": "Llama-3.2-90B-Vision-Instruct",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Llama-3.2-90B-Vision-Instruct-2",
+    "name": "Llama-3.2-90B-Vision-Instruct-2",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Llama-3.2-90B-Vision-Instruct-3",
+    "name": "Llama-3.2-90B-Vision-Instruct-3",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Llama-3.3-70B-Instruct",
+    "name": "Llama-3.3-70B-Instruct",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Llama-3.3-70B-Instruct-2",
+    "name": "Llama-3.3-70B-Instruct-2",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Llama-3.3-70B-Instruct-3",
+    "name": "Llama-3.3-70B-Instruct-3",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Llama-3.3-70B-Instruct-4",
+    "name": "Llama-3.3-70B-Instruct-4",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Llama-3.3-70B-Instruct-5",
+    "name": "Llama-3.3-70B-Instruct-5",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Llama-3.3-70B-Instruct-9",
+    "name": "Llama-3.3-70B-Instruct-9",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Llama-4-Maverick-17B-128E-Instruct-FP8",
+    "name": "Llama-4-Maverick-17B-128E-Instruct-FP8",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Llama-4-Scout-17B-16E-Instruct",
+    "name": "Llama-4-Scout-17B-16E-Instruct",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "MAI-DS-R1",
+    "name": "MAI-DS-R1",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "MAI-Image-2-2026-02-20",
+    "name": "MAI-Image-2-2026-02-20",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "MAI-Image-2e-2026-04-09",
+    "name": "MAI-Image-2e-2026-04-09",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3-70B-Instruct-6",
+    "name": "Meta-Llama-3-70B-Instruct-6",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3-70B-Instruct-7",
+    "name": "Meta-Llama-3-70B-Instruct-7",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3-70B-Instruct-8",
+    "name": "Meta-Llama-3-70B-Instruct-8",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3-70B-Instruct-9",
+    "name": "Meta-Llama-3-70B-Instruct-9",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3-8B-Instruct-6",
+    "name": "Meta-Llama-3-8B-Instruct-6",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3-8B-Instruct-7",
+    "name": "Meta-Llama-3-8B-Instruct-7",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3-8B-Instruct-8",
+    "name": "Meta-Llama-3-8B-Instruct-8",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3-8B-Instruct-9",
+    "name": "Meta-Llama-3-8B-Instruct-9",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3.1-405B-Instruct",
+    "name": "Meta-Llama-3.1-405B-Instruct",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3.1-70B-Instruct",
+    "name": "Meta-Llama-3.1-70B-Instruct",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3.1-70B-Instruct-2",
+    "name": "Meta-Llama-3.1-70B-Instruct-2",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3.1-70B-Instruct-3",
+    "name": "Meta-Llama-3.1-70B-Instruct-3",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3.1-70B-Instruct-4",
+    "name": "Meta-Llama-3.1-70B-Instruct-4",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3.1-8B-Instruct",
+    "name": "Meta-Llama-3.1-8B-Instruct",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3.1-8B-Instruct-2",
+    "name": "Meta-Llama-3.1-8B-Instruct-2",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3.1-8B-Instruct-3",
+    "name": "Meta-Llama-3.1-8B-Instruct-3",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3.1-8B-Instruct-4",
+    "name": "Meta-Llama-3.1-8B-Instruct-4",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Meta-Llama-3.1-8B-Instruct-5",
+    "name": "Meta-Llama-3.1-8B-Instruct-5",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Ministral-3B",
+    "name": "Ministral-3B",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Mistral-Large-2411-2",
+    "name": "Mistral-Large-2411-2",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Mistral-Large-3",
+    "name": "Mistral-Large-3",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Mistral-Nemo",
+    "name": "Mistral-Nemo",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Mistral-large",
+    "name": "Mistral-large",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Mistral-large-2407",
+    "name": "Mistral-large-2407",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Mistral-small",
+    "name": "Mistral-small",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-medium-128k-instruct-3",
+    "name": "Phi-3-medium-128k-instruct-3",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-medium-128k-instruct-4",
+    "name": "Phi-3-medium-128k-instruct-4",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-medium-128k-instruct-5",
+    "name": "Phi-3-medium-128k-instruct-5",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-medium-128k-instruct-6",
+    "name": "Phi-3-medium-128k-instruct-6",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-medium-128k-instruct-7",
+    "name": "Phi-3-medium-128k-instruct-7",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-medium-4k-instruct-3",
+    "name": "Phi-3-medium-4k-instruct-3",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-medium-4k-instruct-4",
+    "name": "Phi-3-medium-4k-instruct-4",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-medium-4k-instruct-5",
+    "name": "Phi-3-medium-4k-instruct-5",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-medium-4k-instruct-6",
+    "name": "Phi-3-medium-4k-instruct-6",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-mini-128k-instruct-10",
+    "name": "Phi-3-mini-128k-instruct-10",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-mini-128k-instruct-11",
+    "name": "Phi-3-mini-128k-instruct-11",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-mini-128k-instruct-12",
+    "name": "Phi-3-mini-128k-instruct-12",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-mini-128k-instruct-13",
+    "name": "Phi-3-mini-128k-instruct-13",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-mini-4k-instruct-10",
+    "name": "Phi-3-mini-4k-instruct-10",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-mini-4k-instruct-11",
+    "name": "Phi-3-mini-4k-instruct-11",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-mini-4k-instruct-13",
+    "name": "Phi-3-mini-4k-instruct-13",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-mini-4k-instruct-14",
+    "name": "Phi-3-mini-4k-instruct-14",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-mini-4k-instruct-15",
+    "name": "Phi-3-mini-4k-instruct-15",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-small-128k-instruct-3",
+    "name": "Phi-3-small-128k-instruct-3",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-small-128k-instruct-4",
+    "name": "Phi-3-small-128k-instruct-4",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-small-128k-instruct-5",
+    "name": "Phi-3-small-128k-instruct-5",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-small-8k-instruct-3",
+    "name": "Phi-3-small-8k-instruct-3",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-small-8k-instruct-4",
+    "name": "Phi-3-small-8k-instruct-4",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3-small-8k-instruct-5",
+    "name": "Phi-3-small-8k-instruct-5",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3.5-MoE-instruct-2",
+    "name": "Phi-3.5-MoE-instruct-2",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3.5-MoE-instruct-3",
+    "name": "Phi-3.5-MoE-instruct-3",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3.5-MoE-instruct-4",
+    "name": "Phi-3.5-MoE-instruct-4",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3.5-MoE-instruct-5",
+    "name": "Phi-3.5-MoE-instruct-5",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3.5-mini-instruct",
+    "name": "Phi-3.5-mini-instruct",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3.5-mini-instruct-2",
+    "name": "Phi-3.5-mini-instruct-2",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3.5-mini-instruct-3",
+    "name": "Phi-3.5-mini-instruct-3",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3.5-mini-instruct-4",
+    "name": "Phi-3.5-mini-instruct-4",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3.5-mini-instruct-6",
+    "name": "Phi-3.5-mini-instruct-6",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3.5-vision-instruct",
+    "name": "Phi-3.5-vision-instruct",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-3.5-vision-instruct-2",
+    "name": "Phi-3.5-vision-instruct-2",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-4-2",
+    "name": "Phi-4-2",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-4-3",
+    "name": "Phi-4-3",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-4-4",
+    "name": "Phi-4-4",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-4-5",
+    "name": "Phi-4-5",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-4-6",
+    "name": "Phi-4-6",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-4-7",
+    "name": "Phi-4-7",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-4-mini-instruct",
+    "name": "Phi-4-mini-instruct",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-4-mini-reasoning",
+    "name": "Phi-4-mini-reasoning",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-4-multimodal-instruct",
+    "name": "Phi-4-multimodal-instruct",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Phi-4-reasoning",
+    "name": "Phi-4-reasoning",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Stable-Diffusion-3.5-Large",
+    "name": "Stable-Diffusion-3.5-Large",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Stable-Image-Core",
+    "name": "Stable-Image-Core",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "Stable-Image-Ultra",
+    "name": "Stable-Image-Ultra",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "ada",
+    "name": "ada",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "aoai-sora",
+    "name": "aoai-sora",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "aoai-sora-2025-02-28",
+    "name": "aoai-sora-2025-02-28",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "babbage",
+    "name": "babbage",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.4,
+          "output_per_million": 0.4
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "claude-haiku-4-5-20251001",
+    "name": "claude-haiku-4-5-20251001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "claude-opus-4-1-20250805",
+    "name": "claude-opus-4-1-20250805",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "claude-opus-4-5-20251101",
+    "name": "claude-opus-4-5-20251101",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "claude-opus-4-6",
+    "name": "claude-opus-4-6",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "claude-opus-4-7",
+    "name": "claude-opus-4-7",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "claude-sonnet-4-5-20250929",
+    "name": "claude-sonnet-4-5-20250929",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "claude-sonnet-4-6",
+    "name": "claude-sonnet-4-6",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "code-cushman-001",
+    "name": "code-cushman-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "code-cushman-fine-tune-002",
+    "name": "code-cushman-fine-tune-002",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "code-davinci-002",
+    "name": "code-davinci-002",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "code-davinci-fine-tune-002",
+    "name": "code-davinci-fine-tune-002",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "code-search-ada-code-001",
+    "name": "code-search-ada-code-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "code-search-ada-text-001",
+    "name": "code-search-ada-text-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "code-search-babbage-code-001",
+    "name": "code-search-babbage-code-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "code-search-babbage-text-001",
+    "name": "code-search-babbage-text-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "codex-mini-2025-05-16",
+    "name": "codex-mini-2025-05-16",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "cohere-command-a",
+    "name": "cohere-command-a",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "computer-use-preview-2025-04-15",
+    "name": "computer-use-preview-2025-04-15",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "curie",
+    "name": "curie",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "dall-e-2",
+    "name": "dall-e-2",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "dall-e-2-2.0",
+    "name": "dall-e-2-2.0",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "dall-e-3",
+    "name": "dall-e-3",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "dall-e-3-3.0",
+    "name": "dall-e-3-3.0",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "davinci",
+    "name": "davinci",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "output_per_million": 2.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "embed-v-4-0",
+    "name": "embed-v-4-0",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-35-turbo",
+    "name": "gpt-35-turbo",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-35-turbo-0125",
+    "name": "gpt-35-turbo-0125",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-35-turbo-0301",
+    "name": "gpt-35-turbo-0301",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-35-turbo-0613",
+    "name": "gpt-35-turbo-0613",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-35-turbo-1106",
+    "name": "gpt-35-turbo-1106",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-35-turbo-16k",
+    "name": "gpt-35-turbo-16k",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-35-turbo-16k-0613",
+    "name": "gpt-35-turbo-16k-0613",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-35-turbo-instruct",
+    "name": "gpt-35-turbo-instruct",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-35-turbo-instruct-0914",
+    "name": "gpt-35-turbo-instruct-0914",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4",
+    "name": "gpt-4",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 8192,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 10.0,
+          "output_per_million": 30.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4-0125-Preview",
+    "name": "gpt-4-0125-Preview",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4-0314",
+    "name": "gpt-4-0314",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4-0613",
+    "name": "gpt-4-0613",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4-1106-Preview",
+    "name": "gpt-4-1106-Preview",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4-32k",
+    "name": "gpt-4-32k",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4-32k-0314",
+    "name": "gpt-4-32k-0314",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4-32k-0613",
+    "name": "gpt-4-32k-0613",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4-turbo-2024-04-09",
+    "name": "gpt-4-turbo-2024-04-09",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 10.0,
+          "output_per_million": 30.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4-turbo-jp",
+    "name": "gpt-4-turbo-jp",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 10.0,
+          "output_per_million": 30.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4-vision-preview",
+    "name": "gpt-4-vision-preview",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4.1",
+    "name": "gpt-4.1",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 1047576,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "output_per_million": 8.0,
+          "cache_read_input_per_million": 0.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4.1-2025-04-14",
+    "name": "gpt-4.1-2025-04-14",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 1047576,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "output_per_million": 8.0,
+          "cache_read_input_per_million": 0.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4.1-2025-04-14-text",
+    "name": "gpt-4.1-2025-04-14-text",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 1047576,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "output_per_million": 8.0,
+          "cache_read_input_per_million": 0.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4.1-mini",
+    "name": "gpt-4.1-mini",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 1047576,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.4,
+          "output_per_million": 1.6,
+          "cache_read_input_per_million": 0.1
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4.1-mini-2025-04-14",
+    "name": "gpt-4.1-mini-2025-04-14",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 1047576,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.4,
+          "output_per_million": 1.6,
+          "cache_read_input_per_million": 0.1
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4.1-nano",
+    "name": "gpt-4.1-nano",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 1047576,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.1,
+          "output_per_million": 0.4
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4.1-nano-2025-04-14",
+    "name": "gpt-4.1-nano-2025-04-14",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 1047576,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.1,
+          "output_per_million": 0.4
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o",
+    "name": "gpt-4o",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-2024-05-13",
+    "name": "gpt-4o-2024-05-13",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-2024-08-06",
+    "name": "gpt-4o-2024-08-06",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-2024-11-20",
+    "name": "gpt-4o-2024-11-20",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-audio-mai",
+    "name": "gpt-4o-audio-mai",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-audio-preview-2024-10-01",
+    "name": "gpt-4o-audio-preview-2024-10-01",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-audio-preview-2024-12-17",
+    "name": "gpt-4o-audio-preview-2024-12-17",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-audio-preview-2025-06-03",
+    "name": "gpt-4o-audio-preview-2025-06-03",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-canvas-2024-09-25",
+    "name": "gpt-4o-canvas-2024-09-25",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-mini",
+    "name": "gpt-4o-mini",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.15,
+          "output_per_million": 0.6
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-mini-2024-07-18",
+    "name": "gpt-4o-mini-2024-07-18",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.15,
+          "output_per_million": 0.6
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-mini-audio-preview-2024-12-17",
+    "name": "gpt-4o-mini-audio-preview-2024-12-17",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.15,
+          "output_per_million": 0.6
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-mini-realtime-preview-2024-12-17",
+    "name": "gpt-4o-mini-realtime-preview-2024-12-17",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.6,
+          "output_per_million": 2.4
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-mini-transcribe",
+    "name": "gpt-4o-mini-transcribe",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 16000,
+    "max_output_tokens": 2000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 5.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-mini-transcribe-2025-03-20",
+    "name": "gpt-4o-mini-transcribe-2025-03-20",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 16000,
+    "max_output_tokens": 2000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 5.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-mini-transcribe-2025-12-15",
+    "name": "gpt-4o-mini-transcribe-2025-12-15",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 16000,
+    "max_output_tokens": 2000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 5.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-mini-tts",
+    "name": "gpt-4o-mini-tts",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.6,
+          "output_per_million": 12.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-mini-tts-2025-03-20",
+    "name": "gpt-4o-mini-tts-2025-03-20",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.6,
+          "output_per_million": 12.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-mini-tts-2025-12-15",
+    "name": "gpt-4o-mini-tts-2025-12-15",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.6,
+          "output_per_million": 12.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-realtime-preview",
+    "name": "gpt-4o-realtime-preview",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5.0,
+          "output_per_million": 20.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-realtime-preview-2024-12-17",
+    "name": "gpt-4o-realtime-preview-2024-12-17",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5.0,
+          "output_per_million": 20.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-realtime-preview-2025-06-03",
+    "name": "gpt-4o-realtime-preview-2025-06-03",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5.0,
+          "output_per_million": 20.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-transcribe",
+    "name": "gpt-4o-transcribe",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-transcribe-2025-03-20",
+    "name": "gpt-4o-transcribe-2025-03-20",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-transcribe-diarize",
+    "name": "gpt-4o-transcribe-diarize",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-4o-transcribe-diarize-2025-10-15",
+    "name": "gpt-4o-transcribe-diarize-2025-10-15",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5-2025-08-07",
+    "name": "gpt-5-2025-08-07",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5-chat-2025-08-07",
+    "name": "gpt-5-chat-2025-08-07",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5-chat-2025-08-15",
+    "name": "gpt-5-chat-2025-08-15",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5-chat-2025-10-03",
+    "name": "gpt-5-chat-2025-10-03",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5-codex-2025-09-15",
+    "name": "gpt-5-codex-2025-09-15",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5-mini-2025-08-07",
+    "name": "gpt-5-mini-2025-08-07",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.25,
+          "output_per_million": 2.0,
+          "cache_read_input_per_million": 0.025
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5-mini-2025-08-07-lite",
+    "name": "gpt-5-mini-2025-08-07-lite",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.25,
+          "output_per_million": 2.0,
+          "cache_read_input_per_million": 0.025
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5-mini-lite-2025-08-07",
+    "name": "gpt-5-mini-lite-2025-08-07",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.25,
+          "output_per_million": 2.0,
+          "cache_read_input_per_million": 0.025
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5-nano-2025-08-07",
+    "name": "gpt-5-nano-2025-08-07",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.05,
+          "output_per_million": 0.4,
+          "cache_read_input_per_million": 0.005
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5-pro-2025-10-06",
+    "name": "gpt-5-pro-2025-10-06",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.1",
+    "name": "gpt-5.1",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.1-2025-11-13",
+    "name": "gpt-5.1-2025-11-13",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.1-chat-2025-11-13",
+    "name": "gpt-5.1-chat-2025-11-13",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.1-codex-2025-11-13",
+    "name": "gpt-5.1-codex-2025-11-13",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.1-codex-max-2025-12-04",
+    "name": "gpt-5.1-codex-max-2025-12-04",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.1-codex-mini-2025-11-13",
+    "name": "gpt-5.1-codex-mini-2025-11-13",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.25,
+          "output_per_million": 2.0,
+          "cache_read_input_per_million": 0.025
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.2-2025-12-11",
+    "name": "gpt-5.2-2025-12-11",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.2-chat-2025-12-11",
+    "name": "gpt-5.2-chat-2025-12-11",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.2-chat-2026-02-10",
+    "name": "gpt-5.2-chat-2026-02-10",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.2-codex-2026-01-14",
+    "name": "gpt-5.2-codex-2026-01-14",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.3-chat-2026-03-03",
+    "name": "gpt-5.3-chat-2026-03-03",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.3-codex-2026-02-20",
+    "name": "gpt-5.3-codex-2026-02-20",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.3-codex-2026-02-24",
+    "name": "gpt-5.3-codex-2026-02-24",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.4-2026-03-05",
+    "name": "gpt-5.4-2026-03-05",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.4-mini-2026-03-17",
+    "name": "gpt-5.4-mini-2026-03-17",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.25,
+          "output_per_million": 2.0,
+          "cache_read_input_per_million": 0.025
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.4-nano-2026-03-17",
+    "name": "gpt-5.4-nano-2026-03-17",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.05,
+          "output_per_million": 0.4,
+          "cache_read_input_per_million": 0.005
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.4-pro-2026-03-05",
+    "name": "gpt-5.4-pro-2026-03-05",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-5.5-2026-04-24",
+    "name": "gpt-5.5-2026-04-24",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-audio-1.5-2026-02-23",
+    "name": "gpt-audio-1.5-2026-02-23",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-audio-2025-08-28",
+    "name": "gpt-audio-2025-08-28",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-audio-mini-2025-10-06",
+    "name": "gpt-audio-mini-2025-10-06",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-chat-latest-2026-05-05",
+    "name": "gpt-chat-latest-2026-05-05",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-image-1",
+    "name": "gpt-image-1",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5.0,
+          "cache_read_input_per_million": 1.25
+        }
+      },
+      "images": {
+        "standard": {
+          "input_per_million": 10.0,
+          "output_per_million": 40.0,
+          "cache_read_input_per_million": 2.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-image-1-2025-04-15",
+    "name": "gpt-image-1-2025-04-15",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5.0,
+          "cache_read_input_per_million": 1.25
+        }
+      },
+      "images": {
+        "standard": {
+          "input_per_million": 10.0,
+          "output_per_million": 40.0,
+          "cache_read_input_per_million": 2.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-image-1-mini",
+    "name": "gpt-image-1-mini",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "cache_read_input_per_million": 0.2
+        }
+      },
+      "images": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 8.0,
+          "cache_read_input_per_million": 0.25
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-image-1-mini-2025-10-06",
+    "name": "gpt-image-1-mini-2025-10-06",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "cache_read_input_per_million": 0.2
+        }
+      },
+      "images": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 8.0,
+          "cache_read_input_per_million": 0.25
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-image-1.5",
+    "name": "gpt-image-1.5",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5.0,
+          "cache_read_input_per_million": 1.25
+        }
+      },
+      "images": {
+        "standard": {
+          "input_per_million": 8.0,
+          "output_per_million": 32.0,
+          "cache_read_input_per_million": 2.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-image-1.5-2025-12-16",
+    "name": "gpt-image-1.5-2025-12-16",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5.0,
+          "cache_read_input_per_million": 1.25
+        }
+      },
+      "images": {
+        "standard": {
+          "input_per_million": 8.0,
+          "output_per_million": 32.0,
+          "cache_read_input_per_million": 2.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-oss-120b",
+    "name": "gpt-oss-120b",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-oss-20b-11",
+    "name": "gpt-oss-20b-11",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-realtime-1.5-2026-02-23",
+    "name": "gpt-realtime-1.5-2026-02-23",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-realtime-2025-08-28",
+    "name": "gpt-realtime-2025-08-28",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-realtime-mini",
+    "name": "gpt-realtime-mini",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-realtime-mini-2025-10-06",
+    "name": "gpt-realtime-mini-2025-10-06",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "gpt-realtime-mini-2025-12-15",
+    "name": "gpt-realtime-mini-2025-12-15",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "grok-3",
+    "name": "grok-3",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "grok-3-mini",
+    "name": "grok-3-mini",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "grok-4-1-fast-non-reasoning",
+    "name": "grok-4-1-fast-non-reasoning",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "grok-4-1-fast-reasoning",
+    "name": "grok-4-1-fast-reasoning",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "grok-4-20-non-reasoning",
+    "name": "grok-4-20-non-reasoning",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "grok-4-20-reasoning",
+    "name": "grok-4-20-reasoning",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "grok-4-fast-non-reasoning",
+    "name": "grok-4-fast-non-reasoning",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "grok-4-fast-reasoning",
+    "name": "grok-4-fast-reasoning",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "jais-30b-chat",
+    "name": "jais-30b-chat",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "jais-30b-chat-2",
+    "name": "jais-30b-chat-2",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "jais-30b-chat-3",
+    "name": "jais-30b-chat-3",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "mistral-document-ai-2505",
+    "name": "mistral-document-ai-2505",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "mistral-document-ai-2512",
+    "name": "mistral-document-ai-2512",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "mistral-medium-2505",
+    "name": "mistral-medium-2505",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "mistral-small-2503",
+    "name": "mistral-small-2503",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "model-router",
+    "name": "model-router",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "model-router-2025-05-19",
+    "name": "model-router-2025-05-19",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "model-router-2025-08-07",
+    "name": "model-router-2025-08-07",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "model-router-2025-11-18",
+    "name": "model-router-2025-11-18",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "o1-2024-12-17",
+    "name": "o1-2024-12-17",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 200000,
+    "max_output_tokens": 100000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 15.0,
+          "output_per_million": 60.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "o1-mini-2024-09-12",
+    "name": "o1-mini-2024-09-12",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.1,
+          "output_per_million": 4.4
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "o1-pro",
+    "name": "o1-pro",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 200000,
+    "max_output_tokens": 100000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 150.0,
+          "output_per_million": 600.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "o1-pro-2025-03-19",
+    "name": "o1-pro-2025-03-19",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 200000,
+    "max_output_tokens": 100000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 150.0,
+          "output_per_million": 600.0
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "o3-deep-research-2025-06-26",
+    "name": "o3-deep-research-2025-06-26",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "o3-deep-research-2025-06-26-ev3",
+    "name": "o3-deep-research-2025-06-26-ev3",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "o3-mini",
+    "name": "o3-mini",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 200000,
+    "max_output_tokens": 100000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.1,
+          "output_per_million": 4.4
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "o3-mini-2025-01-31",
+    "name": "o3-mini-2025-01-31",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 200000,
+    "max_output_tokens": 100000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.1,
+          "output_per_million": 4.4
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "o3-mini-alpha",
+    "name": "o3-mini-alpha",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 200000,
+    "max_output_tokens": 100000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.1,
+          "output_per_million": 4.4
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "o3-mini-alpha-2024-12-17",
+    "name": "o3-mini-alpha-2024-12-17",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 200000,
+    "max_output_tokens": 100000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.1,
+          "output_per_million": 4.4
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "o4-mini",
+    "name": "o4-mini",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "o4-mini-2025-04-16",
+    "name": "o4-mini-2025-04-16",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "qwen-3-32b",
+    "name": "qwen-3-32b",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "qwen3-32b",
+    "name": "qwen3-32b",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "sora",
+    "name": "sora",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "sora-2",
+    "name": "sora-2",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "sora-2-2025-10-06",
+    "name": "sora-2-2025-10-06",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "sora-2-2025-12-08",
+    "name": "sora-2-2025-12-08",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "sora-2025-05-02",
+    "name": "sora-2025-05-02",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-ada-001",
+    "name": "text-ada-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-babbage-001",
+    "name": "text-babbage-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-curie-001",
+    "name": "text-curie-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-davinci-001",
+    "name": "text-davinci-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-davinci-002",
+    "name": "text-davinci-002",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-davinci-003",
+    "name": "text-davinci-003",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-davinci-fine-tune-002",
+    "name": "text-davinci-fine-tune-002",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-embedding-3-large",
+    "name": "text-embedding-3-large",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.13,
+          "output_per_million": 0.13
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-embedding-3-small",
+    "name": "text-embedding-3-small",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.02,
+          "output_per_million": 0.02
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-embedding-ada-002",
+    "name": "text-embedding-ada-002",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.1,
+          "output_per_million": 0.1
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-embedding-ada-002-2",
+    "name": "text-embedding-ada-002-2",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.1,
+          "output_per_million": 0.1
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-search-ada-doc-001",
+    "name": "text-search-ada-doc-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-search-ada-query-001",
+    "name": "text-search-ada-query-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-search-babbage-doc-001",
+    "name": "text-search-babbage-doc-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-search-babbage-query-001",
+    "name": "text-search-babbage-query-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-search-curie-doc-001",
+    "name": "text-search-curie-doc-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-search-curie-query-001",
+    "name": "text-search-curie-query-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-search-davinci-doc-001",
+    "name": "text-search-davinci-doc-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-search-davinci-query-001",
+    "name": "text-search-davinci-query-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-similarity-ada-001",
+    "name": "text-similarity-ada-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-similarity-babbage-001",
+    "name": "text-similarity-babbage-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-similarity-curie-001",
+    "name": "text-similarity-curie-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "text-similarity-davinci-001",
+    "name": "text-similarity-davinci-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "whisper",
+    "name": "whisper",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.006,
+          "output_per_million": 0.006
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
+    }
+  },
+  {
+    "id": "whisper-001",
+    "name": "whisper-001",
+    "provider": "azure",
+    "family": null,
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.006,
+          "output_per_million": 0.006
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": null
     }
   },
   {
@@ -1163,7 +9809,7 @@
     "name": "Nova 2 Lite",
     "provider": "bedrock",
     "family": "nova",
-    "created_at": "2024-12-01 00:00:00 +0530",
+    "created_at": "2024-12-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -1206,12 +9852,44 @@
       }
     }
   },
+  {
+    "id": "amazon.nova-2-sonic-v1:0",
+    "name": "Nova 2 Sonic",
+    "provider": "bedrock",
+    "family": "Nova",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "audio"
+      ],
+      "output": [
+        "audio",
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.nova-2-sonic-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
   {
     "id": "amazon.nova-lite-v1:0",
     "name": "Nova Lite",
     "provider": "bedrock",
     "family": "nova-lite",
-    "created_at": "2024-12-03 00:00:00 +0530",
+    "created_at": "2024-12-03 00:00:00 UTC",
     "context_window": 300000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -1234,7 +9912,7 @@
         "standard": {
           "input_per_million": 0.06,
           "output_per_million": 0.24,
-          "cached_input_per_million": 0.015
+          "cache_read_input_per_million": 0.015
         }
       }
     },
@@ -1262,7 +9940,7 @@
     "name": "Nova Micro",
     "provider": "bedrock",
     "family": "nova-micro",
-    "created_at": "2024-12-03 00:00:00 +0530",
+    "created_at": "2024-12-03 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -1282,7 +9960,7 @@
         "standard": {
           "input_per_million": 0.035,
           "output_per_million": 0.14,
-          "cached_input_per_million": 0.00875
+          "cache_read_input_per_million": 0.00875
         }
       }
     },
@@ -1310,7 +9988,7 @@
     "name": "Nova Premier",
     "provider": "bedrock",
     "family": "nova",
-    "created_at": "2024-12-03 00:00:00 +0530",
+    "created_at": "2024-12-03 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
@@ -1355,12 +10033,208 @@
       "knowledge": "2024-10"
     }
   },
+  {
+    "id": "amazon.nova-premier-v1:0:1000k",
+    "name": "Nova Premier",
+    "provider": "bedrock",
+    "family": "nova",
+    "created_at": "2024-12-03 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 12.5
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.nova-premier-v1:0:1000k",
+      "inference_types": [],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2024-12-03",
+      "cost": {
+        "input": 2.5,
+        "output": 12.5
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 16384
+      },
+      "knowledge": "2024-10"
+    }
+  },
+  {
+    "id": "amazon.nova-premier-v1:0:20k",
+    "name": "Nova Premier",
+    "provider": "bedrock",
+    "family": "nova",
+    "created_at": "2024-12-03 00:00:00 UTC",
+    "context_window": 20000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 12.5
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.nova-premier-v1:0:20k",
+      "inference_types": [],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2024-12-03",
+      "cost": {
+        "input": 2.5,
+        "output": 12.5
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 16384
+      },
+      "knowledge": "2024-10"
+    }
+  },
+  {
+    "id": "amazon.nova-premier-v1:0:8k",
+    "name": "Nova Premier",
+    "provider": "bedrock",
+    "family": "nova",
+    "created_at": "2024-12-03 00:00:00 UTC",
+    "context_window": 8000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 12.5
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.nova-premier-v1:0:8k",
+      "inference_types": [],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2024-12-03",
+      "cost": {
+        "input": 2.5,
+        "output": 12.5
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 16384
+      },
+      "knowledge": "2024-10"
+    }
+  },
+  {
+    "id": "amazon.nova-premier-v1:0:mm",
+    "name": "Nova Premier",
+    "provider": "bedrock",
+    "family": "amazon",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.nova-premier-v1:0:mm",
+      "inference_types": [],
+      "converse": {}
+    }
+  },
   {
     "id": "amazon.nova-pro-v1:0",
     "name": "Nova Pro",
     "provider": "bedrock",
     "family": "nova-pro",
-    "created_at": "2024-12-03 00:00:00 +0530",
+    "created_at": "2024-12-03 00:00:00 UTC",
     "context_window": 300000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -1383,7 +10257,7 @@
         "standard": {
           "input_per_million": 0.8,
           "output_per_million": 3.2,
-          "cached_input_per_million": 0.2
+          "cache_read_input_per_million": 0.2
         }
       }
     },
@@ -1406,15 +10280,259 @@
       "knowledge": "2024-10"
     }
   },
+  {
+    "id": "amazon.rerank-v1:0",
+    "name": "Rerank 1.0",
+    "provider": "bedrock",
+    "family": "amazon",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.rerank-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "amazon.titan-embed-g1-text-02",
+    "name": "Titan Text Embeddings v2",
+    "provider": "bedrock",
+    "family": "amazon",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "embeddings"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.titan-embed-g1-text-02",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "amazon.titan-embed-image-v1",
+    "name": "Titan Multimodal Embeddings G1",
+    "provider": "bedrock",
+    "family": "amazon",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "embeddings"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.titan-embed-image-v1",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "amazon.titan-embed-image-v1:0",
+    "name": "Titan Multimodal Embeddings G1",
+    "provider": "bedrock",
+    "family": "amazon",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "embeddings"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.titan-embed-image-v1:0",
+      "inference_types": [
+        "PROVISIONED"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "amazon.titan-embed-text-v1",
+    "name": "Titan Embeddings G1 - Text",
+    "provider": "bedrock",
+    "family": "amazon",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "embeddings"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.titan-embed-text-v1",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "amazon.titan-embed-text-v1:2:8k",
+    "name": "Titan Embeddings G1 - Text",
+    "provider": "bedrock",
+    "family": "amazon",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "embeddings"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.titan-embed-text-v1:2:8k",
+      "inference_types": [
+        "PROVISIONED"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "amazon.titan-embed-text-v2:0",
+    "name": "Titan Text Embeddings V2",
+    "provider": "bedrock",
+    "family": "amazon",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "embeddings"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.titan-embed-text-v2:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "amazon.titan-image-generator-v2:0",
+    "name": "Titan Image Generator G1 v2",
+    "provider": "bedrock",
+    "family": "amazon",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "image"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.titan-image-generator-v2:0",
+      "inference_types": [
+        "PROVISIONED",
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
   {
     "id": "anthropic.claude-3-5-haiku-20241022-v1:0",
     "name": "Claude Haiku 3.5",
     "provider": "bedrock",
     "family": "claude-haiku",
-    "created_at": "2024-10-22 00:00:00 +0530",
+    "created_at": "2024-10-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 8192,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2024-07-31",
     "modalities": {
       "input": [
         "text",
@@ -1427,18 +10545,26 @@
     },
     "capabilities": [
       "function_calling",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.8,
           "output_per_million": 4,
-          "cached_input_per_million": 0.08
+          "cache_read_input_per_million": 0.08,
+          "cache_write_input_per_million": 1
         }
       }
     },
     "metadata": {
+      "provider_name": "Anthropic",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/anthropic.claude-3-5-haiku-20241022-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {},
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -1455,7 +10581,7 @@
         "context": 200000,
         "output": 8192
       },
-      "knowledge": "2024-07"
+      "knowledge": "2024-07-31"
     }
   },
   {
@@ -1463,10 +10589,10 @@
     "name": "Claude Sonnet 3.5",
     "provider": "bedrock",
     "family": "claude-sonnet",
-    "created_at": "2024-06-20 00:00:00 +0530",
+    "created_at": "2024-06-20 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 8192,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2024-04-30",
     "modalities": {
       "input": [
         "text",
@@ -1486,7 +10612,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -1507,7 +10634,7 @@
         "context": 200000,
         "output": 8192
       },
-      "knowledge": "2024-04"
+      "knowledge": "2024-04-30"
     }
   },
   {
@@ -1515,7 +10642,7 @@
     "name": "Claude Sonnet 3.5 v2",
     "provider": "bedrock",
     "family": "claude-sonnet",
-    "created_at": "2024-10-22 00:00:00 +0530",
+    "created_at": "2024-10-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -1538,7 +10665,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -1567,7 +10695,7 @@
     "name": "Claude Sonnet 3.7",
     "provider": "bedrock",
     "family": "claude-sonnet",
-    "created_at": "2025-02-19 00:00:00 +0530",
+    "created_at": "2025-02-19 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -1590,7 +10718,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -1619,7 +10748,7 @@
     "name": "Claude Haiku 3",
     "provider": "bedrock",
     "family": "claude-haiku",
-    "created_at": "2024-03-13 00:00:00 +0530",
+    "created_at": "2024-03-13 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -1635,7 +10764,8 @@
     },
     "capabilities": [
       "function_calling",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -1646,6 +10776,12 @@
       }
     },
     "metadata": {
+      "provider_name": "Anthropic",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/anthropic.claude-3-haiku-20240307-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {},
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -1663,12 +10799,220 @@
       "knowledge": "2024-02"
     }
   },
+  {
+    "id": "anthropic.claude-3-haiku-20240307-v1:0:200k",
+    "name": "Claude Haiku 3",
+    "provider": "bedrock",
+    "family": "claude-haiku",
+    "created_at": "2024-03-13 00:00:00 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.25,
+          "output_per_million": 1.25
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Anthropic",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/anthropic.claude-3-haiku-20240307-v1:0:200k",
+      "inference_types": [
+        "PROVISIONED"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2024-03-13",
+      "cost": {
+        "input": 0.25,
+        "output": 1.25
+      },
+      "limit": {
+        "context": 200000,
+        "output": 4096
+      },
+      "knowledge": "2024-02"
+    }
+  },
+  {
+    "id": "anthropic.claude-3-haiku-20240307-v1:0:48k",
+    "name": "Claude Haiku 3",
+    "provider": "bedrock",
+    "family": "claude-haiku",
+    "created_at": "2024-03-13 00:00:00 UTC",
+    "context_window": 48000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.25,
+          "output_per_million": 1.25
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Anthropic",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/anthropic.claude-3-haiku-20240307-v1:0:48k",
+      "inference_types": [
+        "PROVISIONED"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2024-03-13",
+      "cost": {
+        "input": 0.25,
+        "output": 1.25
+      },
+      "limit": {
+        "context": 200000,
+        "output": 4096
+      },
+      "knowledge": "2024-02"
+    }
+  },
+  {
+    "id": "anthropic.claude-3-sonnet-20240229-v1:0",
+    "name": "Claude 3 Sonnet",
+    "provider": "bedrock",
+    "family": "anthropic",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Anthropic",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/anthropic.claude-3-sonnet-20240229-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "anthropic.claude-3-sonnet-20240229-v1:0:200k",
+    "name": "Claude 3 Sonnet",
+    "provider": "bedrock",
+    "family": "anthropic",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Anthropic",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/anthropic.claude-3-sonnet-20240229-v1:0:200k",
+      "inference_types": [
+        "PROVISIONED"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "anthropic.claude-3-sonnet-20240229-v1:0:28k",
+    "name": "Claude 3 Sonnet",
+    "provider": "bedrock",
+    "family": "anthropic",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Anthropic",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/anthropic.claude-3-sonnet-20240229-v1:0:28k",
+      "inference_types": [
+        "PROVISIONED"
+      ],
+      "converse": {}
+    }
+  },
   {
     "id": "anthropic.claude-haiku-4-5-20251001-v1:0",
     "name": "Claude Haiku 4.5",
     "provider": "bedrock",
     "family": "claude-haiku",
-    "created_at": "2025-10-15 00:00:00 +0530",
+    "created_at": "2025-10-15 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-02-28",
@@ -1693,7 +11037,8 @@
         "standard": {
           "input_per_million": 1,
           "output_per_million": 5,
-          "cached_input_per_million": 0.1
+          "cache_read_input_per_million": 0.1,
+          "cache_write_input_per_million": 1.25
         }
       }
     },
@@ -1722,7 +11067,7 @@
     "name": "Claude Opus 4.1",
     "provider": "bedrock",
     "family": "claude-opus",
-    "created_at": "2025-08-05 00:00:00 +0530",
+    "created_at": "2025-08-05 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 32000,
     "knowledge_cutoff": "2025-03-31",
@@ -1746,7 +11091,8 @@
         "standard": {
           "input_per_million": 15,
           "output_per_million": 75,
-          "cached_input_per_million": 1.5
+          "cache_read_input_per_million": 1.5,
+          "cache_write_input_per_million": 18.75
         }
       }
     },
@@ -1775,10 +11121,10 @@
     "name": "Claude Opus 4",
     "provider": "bedrock",
     "family": "claude-opus",
-    "created_at": "2025-05-22 00:00:00 +0530",
+    "created_at": "2025-05-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 32000,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2025-03-31",
     "modalities": {
       "input": [
         "text",
@@ -1799,7 +11145,8 @@
         "standard": {
           "input_per_million": 15,
           "output_per_million": 75,
-          "cached_input_per_million": 1.5
+          "cache_read_input_per_million": 1.5,
+          "cache_write_input_per_million": 18.75
         }
       }
     },
@@ -1820,7 +11167,7 @@
         "context": 200000,
         "output": 32000
       },
-      "knowledge": "2024-04"
+      "knowledge": "2025-03-31"
     }
   },
   {
@@ -1828,7 +11175,7 @@
     "name": "Claude Opus 4.5",
     "provider": "bedrock",
     "family": "claude-opus",
-    "created_at": "2025-11-24 00:00:00 +0530",
+    "created_at": "2025-11-24 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-03-31",
@@ -1853,7 +11200,8 @@
         "standard": {
           "input_per_million": 5,
           "output_per_million": 25,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
         }
       }
     },
@@ -1882,10 +11230,10 @@
     "name": "Claude Opus 4.6",
     "provider": "bedrock",
     "family": "claude-opus",
-    "created_at": "2026-02-05 00:00:00 +0530",
+    "created_at": "2026-02-05 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 128000,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2025-05-31",
     "modalities": {
       "input": [
         "text",
@@ -1907,7 +11255,8 @@
         "standard": {
           "input_per_million": 5,
           "output_per_million": 25,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
         }
       }
     },
@@ -1917,7 +11266,7 @@
       "open_weights": false,
       "attachment": true,
       "temperature": true,
-      "last_updated": "2026-03-18",
+      "last_updated": "2026-03-13",
       "cost": {
         "input": 5,
         "output": 25,
@@ -1928,7 +11277,62 @@
         "context": 1000000,
         "output": 128000
       },
-      "knowledge": "2025-05"
+      "knowledge": "2025-05-31"
+    }
+  },
+  {
+    "id": "anthropic.claude-opus-4-7",
+    "name": "Claude Opus 4.7",
+    "provider": "bedrock",
+    "family": "claude-opus",
+    "created_at": "2026-04-16 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": "2026-01-31",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5,
+          "output_per_million": 25,
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": false,
+      "last_updated": "2026-04-16",
+      "cost": {
+        "input": 5,
+        "output": 25,
+        "cache_read": 0.5,
+        "cache_write": 6.25
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 128000
+      },
+      "knowledge": "2026-01-31"
     }
   },
   {
@@ -1936,10 +11340,10 @@
     "name": "Claude Sonnet 4",
     "provider": "bedrock",
     "family": "claude-sonnet",
-    "created_at": "2025-05-22 00:00:00 +0530",
+    "created_at": "2025-05-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2025-03-31",
     "modalities": {
       "input": [
         "text",
@@ -1960,7 +11364,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -1981,7 +11386,7 @@
         "context": 200000,
         "output": 64000
       },
-      "knowledge": "2024-04"
+      "knowledge": "2025-03-31"
     }
   },
   {
@@ -1989,7 +11394,7 @@
     "name": "Claude Sonnet 4.5",
     "provider": "bedrock",
     "family": "claude-sonnet",
-    "created_at": "2025-09-29 00:00:00 +0530",
+    "created_at": "2025-09-29 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-07-31",
@@ -2014,7 +11419,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -2043,10 +11449,10 @@
     "name": "Claude Sonnet 4.6",
     "provider": "bedrock",
     "family": "claude-sonnet",
-    "created_at": "2026-02-17 00:00:00 +0530",
+    "created_at": "2026-02-17 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 64000,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2025-08-31",
     "modalities": {
       "input": [
         "text",
@@ -2067,7 +11473,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -2077,7 +11484,7 @@
       "open_weights": false,
       "attachment": true,
       "temperature": true,
-      "last_updated": "2026-03-18",
+      "last_updated": "2026-03-13",
       "cost": {
         "input": 3,
         "output": 15,
@@ -2088,15 +11495,337 @@
         "context": 1000000,
         "output": 64000
       },
+      "knowledge": "2025-08-31"
+    }
+  },
+  {
+    "id": "au.anthropic.claude-opus-4-6-v1",
+    "name": "AU Anthropic Claude Opus 4.6",
+    "provider": "bedrock",
+    "family": "claude-opus",
+    "created_at": "2026-02-05 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 16.5,
+          "output_per_million": 82.5,
+          "cache_read_input_per_million": 1.65,
+          "cache_write_input_per_million": 20.625
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2026-02-05",
+      "cost": {
+        "input": 16.5,
+        "output": 82.5,
+        "cache_read": 1.65,
+        "cache_write": 20.625
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 128000
+      },
+      "knowledge": "2025-05"
+    }
+  },
+  {
+    "id": "au.anthropic.claude-sonnet-4-6",
+    "name": "AU Anthropic Claude Sonnet 4.6",
+    "provider": "bedrock",
+    "family": "claude-sonnet",
+    "created_at": "2026-02-17 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 3.3,
+          "output_per_million": 16.5,
+          "cache_read_input_per_million": 0.33,
+          "cache_write_input_per_million": 4.125
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2026-02-17",
+      "cost": {
+        "input": 3.3,
+        "output": 16.5,
+        "cache_read": 0.33,
+        "cache_write": 4.125
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 128000
+      },
       "knowledge": "2025-08"
     }
   },
+  {
+    "id": "cohere.command-r-plus-v1:0",
+    "name": "Command R+",
+    "provider": "bedrock",
+    "family": "cohere",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Cohere",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/cohere.command-r-plus-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "cohere.command-r-v1:0",
+    "name": "Command R",
+    "provider": "bedrock",
+    "family": "cohere",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Cohere",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/cohere.command-r-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "cohere.embed-english-v3",
+    "name": "Embed English",
+    "provider": "bedrock",
+    "family": "cohere",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "embeddings"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Cohere",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/cohere.embed-english-v3",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "cohere.embed-english-v3:0:512",
+    "name": "Embed English",
+    "provider": "bedrock",
+    "family": "cohere",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "embeddings"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Cohere",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/cohere.embed-english-v3:0:512",
+      "inference_types": [
+        "PROVISIONED"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "cohere.embed-multilingual-v3",
+    "name": "Embed Multilingual",
+    "provider": "bedrock",
+    "family": "cohere",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "embeddings"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Cohere",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/cohere.embed-multilingual-v3",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "cohere.embed-multilingual-v3:0:512",
+    "name": "Embed Multilingual",
+    "provider": "bedrock",
+    "family": "cohere",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "embeddings"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Cohere",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/cohere.embed-multilingual-v3:0:512",
+      "inference_types": [
+        "PROVISIONED"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "cohere.rerank-v3-5:0",
+    "name": "Rerank 3.5",
+    "provider": "bedrock",
+    "family": "cohere",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Cohere",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/cohere.rerank-v3-5:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
   {
     "id": "deepseek.r1-v1:0",
     "name": "DeepSeek-R1",
     "provider": "bedrock",
     "family": "deepseek-thinking",
-    "created_at": "2025-01-20 00:00:00 +0530",
+    "created_at": "2025-01-20 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -2143,7 +11872,7 @@
     "name": "DeepSeek-V3.1",
     "provider": "bedrock",
     "family": "deepseek",
-    "created_at": "2025-09-18 00:00:00 +0530",
+    "created_at": "2025-09-18 00:00:00 UTC",
     "context_window": 163840,
     "max_output_tokens": 81920,
     "knowledge_cutoff": null,
@@ -2158,7 +11887,8 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -2169,6 +11899,24 @@
       }
     },
     "metadata": {
+      "provider_name": "DeepSeek",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/deepseek.v3-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 163840,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -2191,7 +11939,7 @@
     "name": "DeepSeek-V3.2",
     "provider": "bedrock",
     "family": "deepseek",
-    "created_at": "2026-02-06 00:00:00 +0530",
+    "created_at": "2026-02-06 00:00:00 UTC",
     "context_window": 163840,
     "max_output_tokens": 81920,
     "knowledge_cutoff": null,
@@ -2205,7 +11953,8 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -2216,6 +11965,24 @@
       }
     },
     "metadata": {
+      "provider_name": "DeepSeek",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/deepseek.v3.2",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 163840,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -2238,7 +12005,7 @@
     "name": "Claude Haiku 4.5 (EU)",
     "provider": "bedrock",
     "family": "claude-haiku",
-    "created_at": "2025-10-15 00:00:00 +0530",
+    "created_at": "2025-10-15 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-02-28",
@@ -2263,7 +12030,8 @@
         "standard": {
           "input_per_million": 1,
           "output_per_million": 5,
-          "cached_input_per_million": 0.1
+          "cache_read_input_per_million": 0.1,
+          "cache_write_input_per_million": 1.25
         }
       }
     },
@@ -2292,7 +12060,7 @@
     "name": "Claude Opus 4.5 (EU)",
     "provider": "bedrock",
     "family": "claude-opus",
-    "created_at": "2025-11-24 00:00:00 +0530",
+    "created_at": "2025-11-24 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-03-31",
@@ -2317,7 +12085,8 @@
         "standard": {
           "input_per_million": 5,
           "output_per_million": 25,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
         }
       }
     },
@@ -2346,10 +12115,10 @@
     "name": "Claude Opus 4.6 (EU)",
     "provider": "bedrock",
     "family": "claude-opus",
-    "created_at": "2026-02-05 00:00:00 +0530",
+    "created_at": "2026-02-05 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 128000,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2025-05-31",
     "modalities": {
       "input": [
         "text",
@@ -2371,7 +12140,8 @@
         "standard": {
           "input_per_million": 5,
           "output_per_million": 25,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
         }
       }
     },
@@ -2381,7 +12151,7 @@
       "open_weights": false,
       "attachment": true,
       "temperature": true,
-      "last_updated": "2026-03-18",
+      "last_updated": "2026-03-13",
       "cost": {
         "input": 5,
         "output": 25,
@@ -2392,7 +12162,62 @@
         "context": 1000000,
         "output": 128000
       },
-      "knowledge": "2025-05"
+      "knowledge": "2025-05-31"
+    }
+  },
+  {
+    "id": "eu.anthropic.claude-opus-4-7",
+    "name": "Claude Opus 4.7 (EU)",
+    "provider": "bedrock",
+    "family": "claude-opus",
+    "created_at": "2026-04-16 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": "2026-01-31",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5,
+          "output_per_million": 25,
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": false,
+      "last_updated": "2026-04-16",
+      "cost": {
+        "input": 5,
+        "output": 25,
+        "cache_read": 0.5,
+        "cache_write": 6.25
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 128000
+      },
+      "knowledge": "2026-01-31"
     }
   },
   {
@@ -2400,10 +12225,10 @@
     "name": "Claude Sonnet 4 (EU)",
     "provider": "bedrock",
     "family": "claude-sonnet",
-    "created_at": "2025-05-22 00:00:00 +0530",
+    "created_at": "2025-05-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2025-03-31",
     "modalities": {
       "input": [
         "text",
@@ -2424,7 +12249,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -2445,7 +12271,7 @@
         "context": 200000,
         "output": 64000
       },
-      "knowledge": "2024-04"
+      "knowledge": "2025-03-31"
     }
   },
   {
@@ -2453,7 +12279,7 @@
     "name": "Claude Sonnet 4.5 (EU)",
     "provider": "bedrock",
     "family": "claude-sonnet",
-    "created_at": "2025-09-29 00:00:00 +0530",
+    "created_at": "2025-09-29 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-07-31",
@@ -2478,7 +12304,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -2507,10 +12334,10 @@
     "name": "Claude Sonnet 4.6 (EU)",
     "provider": "bedrock",
     "family": "claude-sonnet",
-    "created_at": "2026-02-17 00:00:00 +0530",
+    "created_at": "2026-02-17 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 64000,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2025-08-31",
     "modalities": {
       "input": [
         "text",
@@ -2531,7 +12358,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -2541,7 +12369,7 @@
       "open_weights": false,
       "attachment": true,
       "temperature": true,
-      "last_updated": "2026-03-18",
+      "last_updated": "2026-03-13",
       "cost": {
         "input": 3,
         "output": 15,
@@ -2552,7 +12380,7 @@
         "context": 1000000,
         "output": 64000
       },
-      "knowledge": "2025-08"
+      "knowledge": "2025-08-31"
     }
   },
   {
@@ -2560,7 +12388,7 @@
     "name": "Claude Haiku 4.5 (Global)",
     "provider": "bedrock",
     "family": "claude-haiku",
-    "created_at": "2025-10-15 00:00:00 +0530",
+    "created_at": "2025-10-15 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-02-28",
@@ -2585,7 +12413,8 @@
         "standard": {
           "input_per_million": 1,
           "output_per_million": 5,
-          "cached_input_per_million": 0.1
+          "cache_read_input_per_million": 0.1,
+          "cache_write_input_per_million": 1.25
         }
       }
     },
@@ -2614,7 +12443,7 @@
     "name": "Claude Opus 4.5 (Global)",
     "provider": "bedrock",
     "family": "claude-opus",
-    "created_at": "2025-11-24 00:00:00 +0530",
+    "created_at": "2025-11-24 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-03-31",
@@ -2639,7 +12468,8 @@
         "standard": {
           "input_per_million": 5,
           "output_per_million": 25,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
         }
       }
     },
@@ -2668,10 +12498,10 @@
     "name": "Claude Opus 4.6 (Global)",
     "provider": "bedrock",
     "family": "claude-opus",
-    "created_at": "2026-02-05 00:00:00 +0530",
+    "created_at": "2026-02-05 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 128000,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2025-05-31",
     "modalities": {
       "input": [
         "text",
@@ -2693,7 +12523,8 @@
         "standard": {
           "input_per_million": 5,
           "output_per_million": 25,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
         }
       }
     },
@@ -2703,7 +12534,7 @@
       "open_weights": false,
       "attachment": true,
       "temperature": true,
-      "last_updated": "2026-03-18",
+      "last_updated": "2026-03-13",
       "cost": {
         "input": 5,
         "output": 25,
@@ -2714,7 +12545,62 @@
         "context": 1000000,
         "output": 128000
       },
-      "knowledge": "2025-05"
+      "knowledge": "2025-05-31"
+    }
+  },
+  {
+    "id": "global.anthropic.claude-opus-4-7",
+    "name": "Claude Opus 4.7 (Global)",
+    "provider": "bedrock",
+    "family": "claude-opus",
+    "created_at": "2026-04-16 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": "2026-01-31",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5,
+          "output_per_million": 25,
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": false,
+      "last_updated": "2026-04-16",
+      "cost": {
+        "input": 5,
+        "output": 25,
+        "cache_read": 0.5,
+        "cache_write": 6.25
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 128000
+      },
+      "knowledge": "2026-01-31"
     }
   },
   {
@@ -2722,10 +12608,10 @@
     "name": "Claude Sonnet 4 (Global)",
     "provider": "bedrock",
     "family": "claude-sonnet",
-    "created_at": "2025-05-22 00:00:00 +0530",
+    "created_at": "2025-05-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2025-03-31",
     "modalities": {
       "input": [
         "text",
@@ -2746,7 +12632,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -2767,7 +12654,7 @@
         "context": 200000,
         "output": 64000
       },
-      "knowledge": "2024-04"
+      "knowledge": "2025-03-31"
     }
   },
   {
@@ -2775,7 +12662,7 @@
     "name": "Claude Sonnet 4.5 (Global)",
     "provider": "bedrock",
     "family": "claude-sonnet",
-    "created_at": "2025-09-29 00:00:00 +0530",
+    "created_at": "2025-09-29 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-07-31",
@@ -2800,7 +12687,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -2829,10 +12717,10 @@
     "name": "Claude Sonnet 4.6 (Global)",
     "provider": "bedrock",
     "family": "claude-sonnet",
-    "created_at": "2026-02-17 00:00:00 +0530",
+    "created_at": "2026-02-17 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 64000,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2025-08-31",
     "modalities": {
       "input": [
         "text",
@@ -2853,7 +12741,8 @@
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
@@ -2863,7 +12752,7 @@
       "open_weights": false,
       "attachment": true,
       "temperature": true,
-      "last_updated": "2026-03-18",
+      "last_updated": "2026-03-13",
       "cost": {
         "input": 3,
         "output": 15,
@@ -2874,7 +12763,7 @@
         "context": 1000000,
         "output": 64000
       },
-      "knowledge": "2025-08"
+      "knowledge": "2025-08-31"
     }
   },
   {
@@ -2882,7 +12771,7 @@
     "name": "Google Gemma 3 12B",
     "provider": "bedrock",
     "family": "gemma",
-    "created_at": "2024-12-01 00:00:00 +0530",
+    "created_at": "2024-12-01 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -2897,7 +12786,9 @@
     },
     "capabilities": [
       "structured_output",
-      "vision"
+      "vision",
+      "streaming",
+      "function_calling"
     ],
     "pricing": {
       "text_tokens": {
@@ -2908,6 +12799,27 @@
       }
     },
     "metadata": {
+      "provider_name": "Google",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/google.gemma-3-12b-it",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 131072,
+        "reasoningSupported": null,
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [
+          "png",
+          "jpeg",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -2930,7 +12842,7 @@
     "name": "Google Gemma 3 27B Instruct",
     "provider": "bedrock",
     "family": "gemma",
-    "created_at": "2025-07-27 00:00:00 +0530",
+    "created_at": "2025-07-27 00:00:00 UTC",
     "context_window": 202752,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -2946,7 +12858,8 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -2957,6 +12870,27 @@
       }
     },
     "metadata": {
+      "provider_name": "Google",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/google.gemma-3-27b-it",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 131072,
+        "reasoningSupported": null,
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [
+          "png",
+          "jpeg",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -2979,7 +12913,7 @@
     "name": "Gemma 3 4B IT",
     "provider": "bedrock",
     "family": "gemma",
-    "created_at": "2024-12-01 00:00:00 +0530",
+    "created_at": "2024-12-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -2994,7 +12928,8 @@
     },
     "capabilities": [
       "function_calling",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -3005,6 +12940,27 @@
       }
     },
     "metadata": {
+      "provider_name": "Google",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/google.gemma-3-4b-it",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 131072,
+        "reasoningSupported": null,
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [
+          "png",
+          "jpeg",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -3021,12 +12977,42 @@
       }
     }
   },
+  {
+    "id": "luma.ray-v2:0",
+    "name": "Ray v2",
+    "provider": "bedrock",
+    "family": "luma ai",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "video"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Luma AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/luma.ray-v2:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
   {
     "id": "meta.llama3-1-405b-instruct-v1:0",
     "name": "Llama 3.1 405B Instruct",
     "provider": "bedrock",
     "family": "llama",
-    "created_at": "2024-07-23 00:00:00 +0530",
+    "created_at": "2024-07-23 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -3039,7 +13025,8 @@
       ]
     },
     "capabilities": [
-      "function_calling"
+      "function_calling",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -3050,6 +13037,12 @@
       }
     },
     "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama3-1-405b-instruct-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {},
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -3072,7 +13065,7 @@
     "name": "Llama 3.1 70B Instruct",
     "provider": "bedrock",
     "family": "llama",
-    "created_at": "2024-07-23 00:00:00 +0530",
+    "created_at": "2024-07-23 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -3085,7 +13078,8 @@
       ]
     },
     "capabilities": [
-      "function_calling"
+      "function_calling",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -3096,6 +13090,66 @@
       }
     },
     "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama3-1-70b-instruct-v1:0",
+      "inference_types": [
+        "ON_DEMAND",
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": true,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2024-07-23",
+      "cost": {
+        "input": 0.72,
+        "output": 0.72
+      },
+      "limit": {
+        "context": 128000,
+        "output": 4096
+      },
+      "knowledge": "2023-12"
+    }
+  },
+  {
+    "id": "meta.llama3-1-70b-instruct-v1:0:128k",
+    "name": "Llama 3.1 70B Instruct",
+    "provider": "bedrock",
+    "family": "llama",
+    "created_at": "2024-07-23 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.72,
+          "output_per_million": 0.72
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama3-1-70b-instruct-v1:0:128k",
+      "inference_types": [
+        "PROVISIONED"
+      ],
+      "converse": {},
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -3118,7 +13172,7 @@
     "name": "Llama 3.1 8B Instruct",
     "provider": "bedrock",
     "family": "llama",
-    "created_at": "2024-07-23 00:00:00 +0530",
+    "created_at": "2024-07-23 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -3131,7 +13185,8 @@
       ]
     },
     "capabilities": [
-      "function_calling"
+      "function_calling",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -3142,6 +13197,66 @@
       }
     },
     "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama3-1-8b-instruct-v1:0",
+      "inference_types": [
+        "ON_DEMAND",
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": true,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2024-07-23",
+      "cost": {
+        "input": 0.22,
+        "output": 0.22
+      },
+      "limit": {
+        "context": 128000,
+        "output": 4096
+      },
+      "knowledge": "2023-12"
+    }
+  },
+  {
+    "id": "meta.llama3-1-8b-instruct-v1:0:128k",
+    "name": "Llama 3.1 8B Instruct",
+    "provider": "bedrock",
+    "family": "llama",
+    "created_at": "2024-07-23 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.22,
+          "output_per_million": 0.22
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama3-1-8b-instruct-v1:0:128k",
+      "inference_types": [
+        "PROVISIONED"
+      ],
+      "converse": {},
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -3164,7 +13279,7 @@
     "name": "Llama 3.2 11B Instruct",
     "provider": "bedrock",
     "family": "llama",
-    "created_at": "2024-09-25 00:00:00 +0530",
+    "created_at": "2024-09-25 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -3207,12 +13322,67 @@
       "knowledge": "2023-12"
     }
   },
+  {
+    "id": "meta.llama3-2-11b-instruct-v1:0:128k",
+    "name": "Llama 3.2 11B Instruct",
+    "provider": "bedrock",
+    "family": "llama",
+    "created_at": "2024-09-25 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.16,
+          "output_per_million": 0.16
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama3-2-11b-instruct-v1:0:128k",
+      "inference_types": [
+        "PROVISIONED"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": true,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2024-09-25",
+      "cost": {
+        "input": 0.16,
+        "output": 0.16
+      },
+      "limit": {
+        "context": 128000,
+        "output": 4096
+      },
+      "knowledge": "2023-12"
+    }
+  },
   {
     "id": "meta.llama3-2-1b-instruct-v1:0",
     "name": "Llama 3.2 1B Instruct",
     "provider": "bedrock",
     "family": "llama",
-    "created_at": "2024-09-25 00:00:00 +0530",
+    "created_at": "2024-09-25 00:00:00 UTC",
     "context_window": 131000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -3253,12 +13423,65 @@
       "knowledge": "2023-12"
     }
   },
+  {
+    "id": "meta.llama3-2-1b-instruct-v1:0:128k",
+    "name": "Llama 3.2 1B Instruct",
+    "provider": "bedrock",
+    "family": "llama",
+    "created_at": "2024-09-25 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.1,
+          "output_per_million": 0.1
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama3-2-1b-instruct-v1:0:128k",
+      "inference_types": [
+        "PROVISIONED"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": true,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2024-09-25",
+      "cost": {
+        "input": 0.1,
+        "output": 0.1
+      },
+      "limit": {
+        "context": 131000,
+        "output": 4096
+      },
+      "knowledge": "2023-12"
+    }
+  },
   {
     "id": "meta.llama3-2-3b-instruct-v1:0",
     "name": "Llama 3.2 3B Instruct",
     "provider": "bedrock",
     "family": "llama",
-    "created_at": "2024-09-25 00:00:00 +0530",
+    "created_at": "2024-09-25 00:00:00 UTC",
     "context_window": 131000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -3299,12 +13522,65 @@
       "knowledge": "2023-12"
     }
   },
+  {
+    "id": "meta.llama3-2-3b-instruct-v1:0:128k",
+    "name": "Llama 3.2 3B Instruct",
+    "provider": "bedrock",
+    "family": "llama",
+    "created_at": "2024-09-25 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.15,
+          "output_per_million": 0.15
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama3-2-3b-instruct-v1:0:128k",
+      "inference_types": [
+        "PROVISIONED"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": true,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2024-09-25",
+      "cost": {
+        "input": 0.15,
+        "output": 0.15
+      },
+      "limit": {
+        "context": 131000,
+        "output": 4096
+      },
+      "knowledge": "2023-12"
+    }
+  },
   {
     "id": "meta.llama3-2-90b-instruct-v1:0",
     "name": "Llama 3.2 90B Instruct",
     "provider": "bedrock",
     "family": "llama",
-    "created_at": "2024-09-25 00:00:00 +0530",
+    "created_at": "2024-09-25 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -3347,12 +13623,67 @@
       "knowledge": "2023-12"
     }
   },
+  {
+    "id": "meta.llama3-2-90b-instruct-v1:0:128k",
+    "name": "Llama 3.2 90B Instruct",
+    "provider": "bedrock",
+    "family": "llama",
+    "created_at": "2024-09-25 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.72,
+          "output_per_million": 0.72
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama3-2-90b-instruct-v1:0:128k",
+      "inference_types": [
+        "PROVISIONED"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": true,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2024-09-25",
+      "cost": {
+        "input": 0.72,
+        "output": 0.72
+      },
+      "limit": {
+        "context": 128000,
+        "output": 4096
+      },
+      "knowledge": "2023-12"
+    }
+  },
   {
     "id": "meta.llama3-3-70b-instruct-v1:0",
     "name": "Llama 3.3 70B Instruct",
     "provider": "bedrock",
     "family": "llama",
-    "created_at": "2024-12-06 00:00:00 +0530",
+    "created_at": "2024-12-06 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -3393,12 +13724,125 @@
       "knowledge": "2023-12"
     }
   },
+  {
+    "id": "meta.llama3-3-70b-instruct-v1:0:128k",
+    "name": "Llama 3.3 70B Instruct",
+    "provider": "bedrock",
+    "family": "llama",
+    "created_at": "2024-12-06 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.72,
+          "output_per_million": 0.72
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama3-3-70b-instruct-v1:0:128k",
+      "inference_types": [],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": true,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2024-12-06",
+      "cost": {
+        "input": 0.72,
+        "output": 0.72
+      },
+      "limit": {
+        "context": 128000,
+        "output": 4096
+      },
+      "knowledge": "2023-12"
+    }
+  },
+  {
+    "id": "meta.llama3-70b-instruct-v1:0",
+    "name": "Llama 3 70B Instruct",
+    "provider": "bedrock",
+    "family": "meta",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama3-70b-instruct-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "meta.llama3-8b-instruct-v1:0",
+    "name": "Llama 3 8B Instruct",
+    "provider": "bedrock",
+    "family": "meta",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama3-8b-instruct-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
   {
     "id": "meta.llama4-maverick-17b-instruct-v1:0",
     "name": "Llama 4 Maverick 17B Instruct",
     "provider": "bedrock",
     "family": "llama",
-    "created_at": "2025-04-05 00:00:00 +0530",
+    "created_at": "2025-04-05 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
@@ -3446,7 +13890,7 @@
     "name": "Llama 4 Scout 17B Instruct",
     "provider": "bedrock",
     "family": "llama",
-    "created_at": "2025-04-05 00:00:00 +0530",
+    "created_at": "2025-04-05 00:00:00 UTC",
     "context_window": 3500000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
@@ -3494,7 +13938,7 @@
     "name": "MiniMax M2",
     "provider": "bedrock",
     "family": "minimax",
-    "created_at": "2025-10-27 00:00:00 +0530",
+    "created_at": "2025-10-27 00:00:00 UTC",
     "context_window": 204608,
     "max_output_tokens": 128000,
     "knowledge_cutoff": null,
@@ -3508,7 +13952,8 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -3519,6 +13964,24 @@
       }
     },
     "metadata": {
+      "provider_name": "MiniMax",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/minimax.minimax-m2",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 409600,
+        "reasoningSupported": {
+          "embedded": true
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -3540,7 +14003,7 @@
     "name": "MiniMax M2.1",
     "provider": "bedrock",
     "family": "minimax",
-    "created_at": "2025-12-23 00:00:00 +0530",
+    "created_at": "2025-12-23 00:00:00 UTC",
     "context_window": 204800,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -3554,7 +14017,8 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -3565,6 +14029,24 @@
       }
     },
     "metadata": {
+      "provider_name": "MiniMax",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/minimax.minimax-m2.1",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 196608,
+        "reasoningSupported": {
+          "embedded": true
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -3586,7 +14068,7 @@
     "name": "MiniMax M2.5",
     "provider": "bedrock",
     "family": "minimax",
-    "created_at": "2026-03-18 00:00:00 +0530",
+    "created_at": "2026-03-18 00:00:00 UTC",
     "context_window": 196608,
     "max_output_tokens": 98304,
     "knowledge_cutoff": null,
@@ -3600,7 +14082,8 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -3611,6 +14094,24 @@
       }
     },
     "metadata": {
+      "provider_name": "MiniMax",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/minimax.minimax-m2.5",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 196608,
+        "reasoningSupported": {
+          "embedded": true
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -3632,7 +14133,7 @@
     "name": "Devstral 2 123B",
     "provider": "bedrock",
     "family": "devstral",
-    "created_at": "2026-02-17 00:00:00 +0530",
+    "created_at": "2026-02-17 00:00:00 UTC",
     "context_window": 256000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -3645,7 +14146,8 @@
       ]
     },
     "capabilities": [
-      "function_calling"
+      "function_calling",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -3656,6 +14158,22 @@
       }
     },
     "metadata": {
+      "provider_name": "Mistral AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/mistral.devstral-2-123b",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 262144,
+        "reasoningSupported": null,
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -3677,7 +14195,7 @@
     "name": "Magistral Small 1.2",
     "provider": "bedrock",
     "family": "magistral",
-    "created_at": "2025-12-02 00:00:00 +0530",
+    "created_at": "2025-12-02 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 40000,
     "knowledge_cutoff": null,
@@ -3694,7 +14212,8 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -3705,6 +14224,29 @@
       }
     },
     "metadata": {
+      "provider_name": "Mistral AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/mistral.magistral-small-2509",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 131072,
+        "reasoningSupported": {
+          "embedded": true
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [
+          "png",
+          "jpeg",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -3726,7 +14268,7 @@
     "name": "Ministral 14B 3.0",
     "provider": "bedrock",
     "family": "ministral",
-    "created_at": "2024-12-01 00:00:00 +0530",
+    "created_at": "2024-12-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -3740,7 +14282,8 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -3751,6 +14294,27 @@
       }
     },
     "metadata": {
+      "provider_name": "Mistral AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/mistral.ministral-3-14b-instruct",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 262144,
+        "reasoningSupported": null,
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [
+          "png",
+          "jpeg",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -3772,7 +14336,7 @@
     "name": "Ministral 3 3B",
     "provider": "bedrock",
     "family": "ministral",
-    "created_at": "2025-12-02 00:00:00 +0530",
+    "created_at": "2025-12-02 00:00:00 UTC",
     "context_window": 256000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -3788,7 +14352,8 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -3799,6 +14364,27 @@
       }
     },
     "metadata": {
+      "provider_name": "Mistral AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/mistral.ministral-3-3b-instruct",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 262144,
+        "reasoningSupported": null,
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [
+          "png",
+          "jpeg",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -3820,7 +14406,7 @@
     "name": "Ministral 3 8B",
     "provider": "bedrock",
     "family": "ministral",
-    "created_at": "2024-12-01 00:00:00 +0530",
+    "created_at": "2024-12-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -3834,7 +14420,8 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -3845,6 +14432,27 @@
       }
     },
     "metadata": {
+      "provider_name": "Mistral AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/mistral.ministral-3-8b-instruct",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 262144,
+        "reasoningSupported": null,
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [
+          "png",
+          "jpeg",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -3861,12 +14469,105 @@
       }
     }
   },
+  {
+    "id": "mistral.mistral-7b-instruct-v0:2",
+    "name": "Mistral 7B Instruct",
+    "provider": "bedrock",
+    "family": "mistral ai",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Mistral AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/mistral.mistral-7b-instruct-v0:2",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "mistral.mistral-large-2402-v1:0",
+    "name": "Mistral Large (24.02)",
+    "provider": "bedrock",
+    "family": "mistral ai",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Mistral AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/mistral.mistral-large-2402-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "mistral.mistral-large-2407-v1:0",
+    "name": "Mistral Large (24.07)",
+    "provider": "bedrock",
+    "family": "mistral ai",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Mistral AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/mistral.mistral-large-2407-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
   {
     "id": "mistral.mistral-large-3-675b-instruct",
     "name": "Mistral Large 3",
     "provider": "bedrock",
     "family": "mistral",
-    "created_at": "2025-12-02 00:00:00 +0530",
+    "created_at": "2025-12-02 00:00:00 UTC",
     "context_window": 256000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -3882,7 +14583,8 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -3893,6 +14595,27 @@
       }
     },
     "metadata": {
+      "provider_name": "Mistral AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/mistral.mistral-large-3-675b-instruct",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 262144,
+        "reasoningSupported": null,
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [
+          "png",
+          "jpeg",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -3909,12 +14632,43 @@
       }
     }
   },
+  {
+    "id": "mistral.mixtral-8x7b-instruct-v0:1",
+    "name": "Mixtral 8x7B Instruct",
+    "provider": "bedrock",
+    "family": "mistral ai",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Mistral AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/mistral.mixtral-8x7b-instruct-v0:1",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
   {
     "id": "mistral.pixtral-large-2502-v1:0",
     "name": "Pixtral Large (25.02)",
     "provider": "bedrock",
     "family": "mistral",
-    "created_at": "2025-04-08 00:00:00 +0530",
+    "created_at": "2025-04-08 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -3961,7 +14715,7 @@
     "name": "Voxtral Mini 3B 2507",
     "provider": "bedrock",
     "family": "mistral",
-    "created_at": "2024-12-01 00:00:00 +0530",
+    "created_at": "2024-12-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -3976,7 +14730,8 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -3987,6 +14742,22 @@
       }
     },
     "metadata": {
+      "provider_name": "Mistral AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/mistral.voxtral-mini-3b-2507",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 32768,
+        "reasoningSupported": null,
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -4008,7 +14779,7 @@
     "name": "Voxtral Small 24B 2507",
     "provider": "bedrock",
     "family": "mistral",
-    "created_at": "2025-07-01 00:00:00 +0530",
+    "created_at": "2025-07-01 00:00:00 UTC",
     "context_window": 32000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -4023,7 +14794,8 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -4034,6 +14806,22 @@
       }
     },
     "metadata": {
+      "provider_name": "Mistral AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/mistral.voxtral-small-24b-2507",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 32768,
+        "reasoningSupported": null,
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -4055,7 +14843,7 @@
     "name": "Kimi K2 Thinking",
     "provider": "bedrock",
     "family": "kimi-thinking",
-    "created_at": "2025-12-02 00:00:00 +0530",
+    "created_at": "2025-12-02 00:00:00 UTC",
     "context_window": 256000,
     "max_output_tokens": 256000,
     "knowledge_cutoff": null,
@@ -4070,7 +14858,8 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -4081,6 +14870,24 @@
       }
     },
     "metadata": {
+      "provider_name": "Moonshot AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/moonshot.kimi-k2-thinking",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 262144,
+        "reasoningSupported": {
+          "embedded": true
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -4103,7 +14910,7 @@
     "name": "Kimi K2.5",
     "provider": "bedrock",
     "family": "kimi",
-    "created_at": "2026-02-06 00:00:00 +0530",
+    "created_at": "2026-02-06 00:00:00 UTC",
     "context_window": 256000,
     "max_output_tokens": 256000,
     "knowledge_cutoff": null,
@@ -4119,7 +14926,8 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -4130,6 +14938,29 @@
       }
     },
     "metadata": {
+      "provider_name": "Moonshot AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/moonshotai.kimi-k2.5",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 262144,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [
+          "jpeg",
+          "png",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -4152,7 +14983,7 @@
     "name": "NVIDIA Nemotron Nano 12B v2 VL BF16",
     "provider": "bedrock",
     "family": "nemotron",
-    "created_at": "2024-12-01 00:00:00 +0530",
+    "created_at": "2024-12-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -4168,7 +14999,8 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -4179,6 +15011,27 @@
       }
     },
     "metadata": {
+      "provider_name": "NVIDIA",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/nvidia.nemotron-nano-12b-v2",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 131072,
+        "reasoningSupported": null,
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [
+          "png",
+          "jpeg",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -4200,7 +15053,7 @@
     "name": "NVIDIA Nemotron Nano 3 30B",
     "provider": "bedrock",
     "family": "nemotron",
-    "created_at": "2025-12-23 00:00:00 +0530",
+    "created_at": "2025-12-23 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -4214,7 +15067,8 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -4225,6 +15079,24 @@
       }
     },
     "metadata": {
+      "provider_name": "NVIDIA",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/nvidia.nemotron-nano-3-30b",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 262144,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -4246,7 +15118,7 @@
     "name": "NVIDIA Nemotron Nano 9B v2",
     "provider": "bedrock",
     "family": "nemotron",
-    "created_at": "2024-12-01 00:00:00 +0530",
+    "created_at": "2024-12-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -4260,7 +15132,8 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -4271,6 +15144,22 @@
       }
     },
     "metadata": {
+      "provider_name": "NVIDIA",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/nvidia.nemotron-nano-9b-v2",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 131072,
+        "reasoningSupported": null,
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -4292,7 +15181,7 @@
     "name": "NVIDIA Nemotron 3 Super 120B A12B",
     "provider": "bedrock",
     "family": "nemotron",
-    "created_at": "2026-03-11 00:00:00 +0530",
+    "created_at": "2026-03-11 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -4306,7 +15195,8 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -4317,6 +15207,22 @@
       }
     },
     "metadata": {
+      "provider_name": "NVIDIA",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/nvidia.nemotron-super-3-120b",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 262144,
+        "reasoningSupported": null,
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -4338,7 +15244,7 @@
     "name": "gpt-oss-120b",
     "provider": "bedrock",
     "family": "gpt-oss",
-    "created_at": "2024-12-01 00:00:00 +0530",
+    "created_at": "2024-12-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -4352,7 +15258,8 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -4363,6 +15270,24 @@
       }
     },
     "metadata": {
+      "provider_name": "OpenAI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/openai.gpt-oss-120b-1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": null,
+        "maxTokensDefault": 4096,
+        "maxTokensMaximum": 128000,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": null,
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -4384,7 +15309,7 @@
     "name": "gpt-oss-20b",
     "provider": "bedrock",
     "family": "gpt-oss",
-    "created_at": "2024-12-01 00:00:00 +0530",
+    "created_at": "2024-12-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -4398,7 +15323,8 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -4409,6 +15335,24 @@
       }
     },
     "metadata": {
+      "provider_name": "OpenAI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/openai.gpt-oss-20b-1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": null,
+        "maxTokensDefault": 4096,
+        "maxTokensMaximum": 128000,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": null,
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -4430,7 +15374,7 @@
     "name": "GPT OSS Safeguard 120B",
     "provider": "bedrock",
     "family": "gpt-oss",
-    "created_at": "2024-12-01 00:00:00 +0530",
+    "created_at": "2024-12-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -4444,7 +15388,9 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming",
+      "reasoning"
     ],
     "pricing": {
       "text_tokens": {
@@ -4455,6 +15401,24 @@
       }
     },
     "metadata": {
+      "provider_name": "OpenAI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/openai.gpt-oss-safeguard-120b",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 131072,
+        "reasoningSupported": {
+          "embedded": true
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -4476,7 +15440,7 @@
     "name": "GPT OSS Safeguard 20B",
     "provider": "bedrock",
     "family": "gpt-oss",
-    "created_at": "2024-12-01 00:00:00 +0530",
+    "created_at": "2024-12-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -4490,7 +15454,9 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming",
+      "reasoning"
     ],
     "pricing": {
       "text_tokens": {
@@ -4501,6 +15467,24 @@
       }
     },
     "metadata": {
+      "provider_name": "OpenAI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/openai.gpt-oss-safeguard-20b",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 131072,
+        "reasoningSupported": {
+          "embedded": true
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -4522,7 +15506,7 @@
     "name": "Qwen3 235B A22B 2507",
     "provider": "bedrock",
     "family": "qwen",
-    "created_at": "2025-09-18 00:00:00 +0530",
+    "created_at": "2025-09-18 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -4536,7 +15520,8 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -4547,6 +15532,22 @@
       }
     },
     "metadata": {
+      "provider_name": "Qwen",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/qwen.qwen3-235b-a22b-2507-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 262144,
+        "reasoningSupported": null,
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -4569,7 +15570,7 @@
     "name": "Qwen3 32B (dense)",
     "provider": "bedrock",
     "family": "qwen",
-    "created_at": "2025-09-18 00:00:00 +0530",
+    "created_at": "2025-09-18 00:00:00 UTC",
     "context_window": 16384,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
@@ -4584,7 +15585,8 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -4595,6 +15597,24 @@
       }
     },
     "metadata": {
+      "provider_name": "Qwen",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/qwen.qwen3-32b-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":true}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 32768,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -4617,7 +15637,7 @@
     "name": "Qwen3 Coder 30B A3B Instruct",
     "provider": "bedrock",
     "family": "qwen",
-    "created_at": "2025-09-18 00:00:00 +0530",
+    "created_at": "2025-09-18 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -4631,7 +15651,8 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -4642,6 +15663,22 @@
       }
     },
     "metadata": {
+      "provider_name": "Qwen",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/qwen.qwen3-coder-30b-a3b-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 262144,
+        "reasoningSupported": null,
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -4664,7 +15701,7 @@
     "name": "Qwen3 Coder 480B A35B Instruct",
     "provider": "bedrock",
     "family": "qwen",
-    "created_at": "2025-09-18 00:00:00 +0530",
+    "created_at": "2025-09-18 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -4678,7 +15715,8 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -4689,6 +15727,22 @@
       }
     },
     "metadata": {
+      "provider_name": "Qwen",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/qwen.qwen3-coder-480b-a35b-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 131072,
+        "reasoningSupported": null,
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -4711,7 +15765,7 @@
     "name": "Qwen3 Coder Next",
     "provider": "bedrock",
     "family": "qwen",
-    "created_at": "2026-02-06 00:00:00 +0530",
+    "created_at": "2026-02-06 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -4758,7 +15812,7 @@
     "name": "Qwen/Qwen3-Next-80B-A3B-Instruct",
     "provider": "bedrock",
     "family": "qwen",
-    "created_at": "2025-09-18 00:00:00 +0530",
+    "created_at": "2025-09-18 00:00:00 UTC",
     "context_window": 262000,
     "max_output_tokens": 262000,
     "knowledge_cutoff": null,
@@ -4772,7 +15826,8 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -4783,6 +15838,24 @@
       }
     },
     "metadata": {
+      "provider_name": "Qwen",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/qwen.qwen3-next-80b-a3b",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 262144,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -4804,7 +15877,7 @@
     "name": "Qwen/Qwen3-VL-235B-A22B-Instruct",
     "provider": "bedrock",
     "family": "qwen",
-    "created_at": "2025-10-04 00:00:00 +0530",
+    "created_at": "2025-10-04 00:00:00 UTC",
     "context_window": 262000,
     "max_output_tokens": 262000,
     "knowledge_cutoff": null,
@@ -4820,7 +15893,8 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -4831,6 +15905,29 @@
       }
     },
     "metadata": {
+      "provider_name": "Qwen",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/qwen.qwen3-vl-235b-a22b",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 262144,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [
+          "png",
+          "jpeg",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -4847,12 +15944,444 @@
       }
     }
   },
+  {
+    "id": "stability.sd3-5-large-v1:0",
+    "name": "Stable Diffusion 3.5 Large",
+    "provider": "bedrock",
+    "family": "stability ai",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "image"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Stability AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/stability.sd3-5-large-v1:0",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "stability.stable-image-core-v1:1",
+    "name": "Stable Image Core 1.0",
+    "provider": "bedrock",
+    "family": "stability ai",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "image"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Stability AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/stability.stable-image-core-v1:1",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "stability.stable-image-ultra-v1:1",
+    "name": "Stable Image Ultra 1.0",
+    "provider": "bedrock",
+    "family": "stability ai",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "image"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Stability AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/stability.stable-image-ultra-v1:1",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "us.amazon.nova-2-lite-v1:0",
+    "name": "Nova 2 Lite",
+    "provider": "bedrock",
+    "family": "nova",
+    "created_at": "2024-12-01 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.33,
+          "output_per_million": 2.75
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.nova-2-lite-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"topK\":{\"type\":\"number\",\"default\":50,\"minimum\":1,\"maximum\":100},\"reasoningConfig\":{\"budgetTokens\":{\"type\":\"enum\",\"default\":\"low\",\"minimum\":10000,\"maximum\":64000,\"enum\":{\"low\":10000,\"medium\":40000,\"high\":64000},\"smartSyncEnabled\":true}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 65535,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": null,
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [
+          "pdf",
+          "docx"
+        ],
+        "userImageTypesSupported": [
+          "jpeg",
+          "png",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": [
+          "mkv",
+          "mov",
+          "mp4",
+          "webm",
+          "flv",
+          "mpeg",
+          "mpg",
+          "wmv",
+          "three_gp"
+        ]
+      },
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2024-12-01",
+      "cost": {
+        "input": 0.33,
+        "output": 2.75
+      },
+      "limit": {
+        "context": 128000,
+        "output": 4096
+      }
+    }
+  },
+  {
+    "id": "us.amazon.nova-lite-v1:0",
+    "name": "Nova Lite",
+    "provider": "bedrock",
+    "family": "nova-lite",
+    "created_at": "2024-12-03 00:00:00 UTC",
+    "context_window": 300000,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.06,
+          "output_per_million": 0.24,
+          "cache_read_input_per_million": 0.015
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.nova-lite-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2024-12-03",
+      "cost": {
+        "input": 0.06,
+        "output": 0.24,
+        "cache_read": 0.015
+      },
+      "limit": {
+        "context": 300000,
+        "output": 8192
+      },
+      "knowledge": "2024-10"
+    }
+  },
+  {
+    "id": "us.amazon.nova-micro-v1:0",
+    "name": "Nova Micro",
+    "provider": "bedrock",
+    "family": "nova-micro",
+    "created_at": "2024-12-03 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.035,
+          "output_per_million": 0.14,
+          "cache_read_input_per_million": 0.00875
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.nova-micro-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2024-12-03",
+      "cost": {
+        "input": 0.035,
+        "output": 0.14,
+        "cache_read": 0.00875
+      },
+      "limit": {
+        "context": 128000,
+        "output": 8192
+      },
+      "knowledge": "2024-10"
+    }
+  },
+  {
+    "id": "us.amazon.nova-premier-v1:0",
+    "name": "Nova Premier",
+    "provider": "bedrock",
+    "family": "nova",
+    "created_at": "2024-12-03 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 12.5
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.nova-premier-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2024-12-03",
+      "cost": {
+        "input": 2.5,
+        "output": 12.5
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 16384
+      },
+      "knowledge": "2024-10"
+    }
+  },
+  {
+    "id": "us.amazon.nova-pro-v1:0",
+    "name": "Nova Pro",
+    "provider": "bedrock",
+    "family": "nova-pro",
+    "created_at": "2024-12-03 00:00:00 UTC",
+    "context_window": 300000,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.8,
+          "output_per_million": 3.2,
+          "cache_read_input_per_million": 0.2
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Amazon",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/amazon.nova-pro-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": null,
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 10000,
+        "reasoningSupported": null,
+        "stopSequencesDefault": null,
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [
+          "pdf",
+          "docx"
+        ],
+        "userImageTypesSupported": [
+          "jpeg",
+          "png",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": [
+          "mkv",
+          "mov",
+          "mp4",
+          "webm",
+          "flv",
+          "mpeg",
+          "mpg",
+          "wmv",
+          "three_gp"
+        ]
+      },
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2024-12-03",
+      "cost": {
+        "input": 0.8,
+        "output": 3.2,
+        "cache_read": 0.2
+      },
+      "limit": {
+        "context": 300000,
+        "output": 8192
+      },
+      "knowledge": "2024-10"
+    }
+  },
   {
     "id": "us.anthropic.claude-haiku-4-5-20251001-v1:0",
     "name": "Claude Haiku 4.5 (US)",
     "provider": "bedrock",
     "family": "claude-haiku",
-    "created_at": "2025-10-15 00:00:00 +0530",
+    "created_at": "2025-10-15 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-02-28",
@@ -4870,18 +16399,45 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1,
           "output_per_million": 5,
-          "cached_input_per_million": 0.1
+          "cache_read_input_per_million": 0.1,
+          "cache_write_input_per_million": 1.25
         }
       }
     },
     "metadata": {
+      "provider_name": "Anthropic",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/anthropic.claude-haiku-4-5-20251001-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"top_k\":{\"type\":\"number\",\"default\":250,\"minimum\":0,\"maximum\":500},\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false},\"budgetTokens\":{\"type\":\"integer\",\"default\":2048,\"minimum\":1024,\"maximum\":63999}}}",
+        "maxTokensDefault": 64000,
+        "maxTokensMaximum": 64000,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": null,
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [
+          "pdf"
+        ],
+        "userImageTypesSupported": [
+          "jpeg",
+          "png",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -4906,7 +16462,7 @@
     "name": "Claude Opus 4.1 (US)",
     "provider": "bedrock",
     "family": "claude-opus",
-    "created_at": "2025-08-05 00:00:00 +0530",
+    "created_at": "2025-08-05 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 32000,
     "knowledge_cutoff": "2025-03-31",
@@ -4923,18 +16479,45 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 15,
           "output_per_million": 75,
-          "cached_input_per_million": 1.5
+          "cache_read_input_per_million": 1.5,
+          "cache_write_input_per_million": 18.75
         }
       }
     },
     "metadata": {
+      "provider_name": "Anthropic",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/anthropic.claude-opus-4-1-20250805-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"topK\":{\"type\":\"number\",\"default\":250,\"minimum\":0,\"maximum\":500},\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false},\"budgetTokens\":{\"type\":\"integer\",\"default\":2048,\"minimum\":1024,\"maximum\":63999}}}",
+        "maxTokensDefault": 1024,
+        "maxTokensMaximum": 32000,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": null,
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [
+          "pdf"
+        ],
+        "userImageTypesSupported": [
+          "jpeg",
+          "png",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -4959,10 +16542,10 @@
     "name": "Claude Opus 4 (US)",
     "provider": "bedrock",
     "family": "claude-opus",
-    "created_at": "2025-05-22 00:00:00 +0530",
+    "created_at": "2025-05-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 32000,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2025-03-31",
     "modalities": {
       "input": [
         "text",
@@ -4976,18 +16559,26 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 15,
           "output_per_million": 75,
-          "cached_input_per_million": 1.5
+          "cache_read_input_per_million": 1.5,
+          "cache_write_input_per_million": 18.75
         }
       }
     },
     "metadata": {
+      "provider_name": "Anthropic",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/anthropic.claude-opus-4-20250514-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {},
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -5004,7 +16595,7 @@
         "context": 200000,
         "output": 32000
       },
-      "knowledge": "2024-04"
+      "knowledge": "2025-03-31"
     }
   },
   {
@@ -5012,7 +16603,7 @@
     "name": "Claude Opus 4.5 (US)",
     "provider": "bedrock",
     "family": "claude-opus",
-    "created_at": "2025-11-24 00:00:00 +0530",
+    "created_at": "2025-11-24 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-03-31",
@@ -5030,18 +16621,45 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 5,
           "output_per_million": 25,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
         }
       }
     },
     "metadata": {
+      "provider_name": "Anthropic",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/anthropic.claude-opus-4-5-20251101-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"top_k\":{\"type\":\"number\",\"default\":250,\"minimum\":0,\"maximum\":500},\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false},\"budgetTokens\":{\"type\":\"integer\",\"default\":2048,\"minimum\":1024,\"maximum\":63999}}}",
+        "maxTokensDefault": 64000,
+        "maxTokensMaximum": 64000,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [
+          "pdf"
+        ],
+        "userImageTypesSupported": [
+          "jpeg",
+          "png",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -5066,10 +16684,10 @@
     "name": "Claude Opus 4.6 (US)",
     "provider": "bedrock",
     "family": "claude-opus",
-    "created_at": "2026-02-05 00:00:00 +0530",
+    "created_at": "2026-02-05 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 128000,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2025-05-31",
     "modalities": {
       "input": [
         "text",
@@ -5084,24 +16702,51 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 5,
           "output_per_million": 25,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
         }
       }
     },
     "metadata": {
+      "provider_name": "Anthropic",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/anthropic.claude-opus-4-6-v1",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"top_k\":{\"type\":\"number\",\"default\":250,\"minimum\":0,\"maximum\":500},\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false},\"budgetTokens\":{\"type\":\"enum\",\"default\":\"low\",\"enum\":{\"low\":1024,\"medium\":40000,\"high\":63999},\"minimum\":1024,\"maximum\":63999}}}",
+        "maxTokensDefault": 128000,
+        "maxTokensMaximum": 128000,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [
+          "pdf"
+        ],
+        "userImageTypesSupported": [
+          "jpeg",
+          "png",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
       "attachment": true,
       "temperature": true,
-      "last_updated": "2026-03-18",
+      "last_updated": "2026-03-13",
       "cost": {
         "input": 5,
         "output": 25,
@@ -5112,7 +16757,88 @@
         "context": 1000000,
         "output": 128000
       },
-      "knowledge": "2025-05"
+      "knowledge": "2025-05-31"
+    }
+  },
+  {
+    "id": "us.anthropic.claude-opus-4-7",
+    "name": "Claude Opus 4.7 (US)",
+    "provider": "bedrock",
+    "family": "claude-opus",
+    "created_at": "2026-04-16 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": "2026-01-31",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5,
+          "output_per_million": 25,
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Anthropic",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/anthropic.claude-opus-4-7",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false},\"budgetTokens\":{\"type\":\"enum\",\"default\":\"low\",\"enum\":{\"low\":1024,\"medium\":40000,\"high\":63999},\"minimum\":1024,\"maximum\":63999}},\"hideSamplingParameter\":true}",
+        "maxTokensDefault": 4096,
+        "maxTokensMaximum": 128000,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [
+          "pdf"
+        ],
+        "userImageTypesSupported": [
+          "jpeg",
+          "png",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": false,
+      "last_updated": "2026-04-16",
+      "cost": {
+        "input": 5,
+        "output": 25,
+        "cache_read": 0.5,
+        "cache_write": 6.25
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 128000
+      },
+      "knowledge": "2026-01-31"
     }
   },
   {
@@ -5120,10 +16846,10 @@
     "name": "Claude Sonnet 4 (US)",
     "provider": "bedrock",
     "family": "claude-sonnet",
-    "created_at": "2025-05-22 00:00:00 +0530",
+    "created_at": "2025-05-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2025-03-31",
     "modalities": {
       "input": [
         "text",
@@ -5137,18 +16863,45 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
     "metadata": {
+      "provider_name": "Anthropic",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/anthropic.claude-sonnet-4-20250514-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": null,
+        "maxTokensDefault": 8192,
+        "maxTokensMaximum": 65536,
+        "reasoningSupported": {
+          "embedded": true
+        },
+        "stopSequencesDefault": null,
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [
+          "pdf"
+        ],
+        "userImageTypesSupported": [
+          "jpeg",
+          "png",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -5165,7 +16918,7 @@
         "context": 200000,
         "output": 64000
       },
-      "knowledge": "2024-04"
+      "knowledge": "2025-03-31"
     }
   },
   {
@@ -5173,7 +16926,7 @@
     "name": "Claude Sonnet 4.5 (US)",
     "provider": "bedrock",
     "family": "claude-sonnet",
-    "created_at": "2025-09-29 00:00:00 +0530",
+    "created_at": "2025-09-29 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-07-31",
@@ -5191,18 +16944,45 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
     "metadata": {
+      "provider_name": "Anthropic",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/anthropic.claude-sonnet-4-5-20250929-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"top_k\":{\"type\":\"number\",\"default\":250,\"minimum\":0,\"maximum\":500},\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false},\"budgetTokens\":{\"type\":\"integer\",\"default\":2048,\"minimum\":1024,\"maximum\":63999}}}",
+        "maxTokensDefault": 64000,
+        "maxTokensMaximum": 64000,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": null,
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [
+          "pdf"
+        ],
+        "userImageTypesSupported": [
+          "jpeg",
+          "png",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
@@ -5227,10 +17007,10 @@
     "name": "Claude Sonnet 4.6 (US)",
     "provider": "bedrock",
     "family": "claude-sonnet",
-    "created_at": "2026-02-17 00:00:00 +0530",
+    "created_at": "2026-02-17 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 64000,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2025-08-31",
     "modalities": {
       "input": [
         "text",
@@ -5244,24 +17024,52 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming",
+      "structured_output"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
     "metadata": {
+      "provider_name": "Anthropic",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/anthropic.claude-sonnet-4-6",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"top_k\":{\"type\":\"number\",\"default\":250,\"minimum\":0,\"maximum\":500},\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false},\"budgetTokens\":{\"type\":\"enum\",\"default\":\"low\",\"enum\":{\"low\":1024,\"medium\":40000,\"high\":63999},\"minimum\":1024,\"maximum\":63999}}}",
+        "maxTokensDefault": 128000,
+        "maxTokensMaximum": 128000,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [
+          "pdf"
+        ],
+        "userImageTypesSupported": [
+          "jpeg",
+          "png",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": false,
       "attachment": true,
       "temperature": true,
-      "last_updated": "2026-03-18",
+      "last_updated": "2026-03-13",
       "cost": {
         "input": 3,
         "output": 15,
@@ -5272,7 +17080,1113 @@
         "context": 1000000,
         "output": 64000
       },
-      "knowledge": "2025-08"
+      "knowledge": "2025-08-31"
+    }
+  },
+  {
+    "id": "us.cohere.embed-v4:0",
+    "name": "Embed v4",
+    "provider": "bedrock",
+    "family": "Embed",
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "embeddings"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Cohere",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/cohere.embed-v4:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "us.deepseek.r1-v1:0",
+    "name": "DeepSeek-R1",
+    "provider": "bedrock",
+    "family": "deepseek-thinking",
+    "created_at": "2025-01-20 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.35,
+          "output_per_million": 5.4
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "DeepSeek",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/deepseek.r1-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2025-05-29",
+      "cost": {
+        "input": 1.35,
+        "output": 5.4
+      },
+      "limit": {
+        "context": 128000,
+        "output": 32768
+      },
+      "knowledge": "2024-07"
+    }
+  },
+  {
+    "id": "us.meta.llama3-2-11b-instruct-v1:0",
+    "name": "Llama 3.2 11B Instruct",
+    "provider": "bedrock",
+    "family": "llama",
+    "created_at": "2024-09-25 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.16,
+          "output_per_million": 0.16
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama3-2-11b-instruct-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": true,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2024-09-25",
+      "cost": {
+        "input": 0.16,
+        "output": 0.16
+      },
+      "limit": {
+        "context": 128000,
+        "output": 4096
+      },
+      "knowledge": "2023-12"
+    }
+  },
+  {
+    "id": "us.meta.llama3-2-1b-instruct-v1:0",
+    "name": "Llama 3.2 1B Instruct",
+    "provider": "bedrock",
+    "family": "llama",
+    "created_at": "2024-09-25 00:00:00 UTC",
+    "context_window": 131000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.1,
+          "output_per_million": 0.1
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama3-2-1b-instruct-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": true,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2024-09-25",
+      "cost": {
+        "input": 0.1,
+        "output": 0.1
+      },
+      "limit": {
+        "context": 131000,
+        "output": 4096
+      },
+      "knowledge": "2023-12"
+    }
+  },
+  {
+    "id": "us.meta.llama3-2-3b-instruct-v1:0",
+    "name": "Llama 3.2 3B Instruct",
+    "provider": "bedrock",
+    "family": "llama",
+    "created_at": "2024-09-25 00:00:00 UTC",
+    "context_window": 131000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.15,
+          "output_per_million": 0.15
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama3-2-3b-instruct-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": true,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2024-09-25",
+      "cost": {
+        "input": 0.15,
+        "output": 0.15
+      },
+      "limit": {
+        "context": 131000,
+        "output": 4096
+      },
+      "knowledge": "2023-12"
+    }
+  },
+  {
+    "id": "us.meta.llama3-2-90b-instruct-v1:0",
+    "name": "Llama 3.2 90B Instruct",
+    "provider": "bedrock",
+    "family": "llama",
+    "created_at": "2024-09-25 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.72,
+          "output_per_million": 0.72
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama3-2-90b-instruct-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": true,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2024-09-25",
+      "cost": {
+        "input": 0.72,
+        "output": 0.72
+      },
+      "limit": {
+        "context": 128000,
+        "output": 4096
+      },
+      "knowledge": "2023-12"
+    }
+  },
+  {
+    "id": "us.meta.llama3-3-70b-instruct-v1:0",
+    "name": "Llama 3.3 70B Instruct",
+    "provider": "bedrock",
+    "family": "llama",
+    "created_at": "2024-12-06 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.72,
+          "output_per_million": 0.72
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama3-3-70b-instruct-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": true,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2024-12-06",
+      "cost": {
+        "input": 0.72,
+        "output": 0.72
+      },
+      "limit": {
+        "context": 128000,
+        "output": 4096
+      },
+      "knowledge": "2023-12"
+    }
+  },
+  {
+    "id": "us.meta.llama4-maverick-17b-instruct-v1:0",
+    "name": "Llama 4 Maverick 17B Instruct",
+    "provider": "bedrock",
+    "family": "llama",
+    "created_at": "2025-04-05 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.24,
+          "output_per_million": 0.97
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama4-maverick-17b-instruct-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": true,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2025-04-05",
+      "cost": {
+        "input": 0.24,
+        "output": 0.97
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 16384
+      },
+      "knowledge": "2024-08"
+    }
+  },
+  {
+    "id": "us.meta.llama4-scout-17b-instruct-v1:0",
+    "name": "Llama 4 Scout 17B Instruct",
+    "provider": "bedrock",
+    "family": "llama",
+    "created_at": "2025-04-05 00:00:00 UTC",
+    "context_window": 3500000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.17,
+          "output_per_million": 0.66
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Meta",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/meta.llama4-scout-17b-instruct-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": true,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2025-04-05",
+      "cost": {
+        "input": 0.17,
+        "output": 0.66
+      },
+      "limit": {
+        "context": 3500000,
+        "output": 16384
+      },
+      "knowledge": "2024-08"
+    }
+  },
+  {
+    "id": "us.mistral.pixtral-large-2502-v1:0",
+    "name": "Pixtral Large (25.02)",
+    "provider": "bedrock",
+    "family": "mistral",
+    "created_at": "2025-04-08 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2,
+          "output_per_million": 6
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Mistral AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/mistral.pixtral-large-2502-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2025-04-08",
+      "cost": {
+        "input": 2,
+        "output": 6
+      },
+      "limit": {
+        "context": 128000,
+        "output": 8192
+      }
+    }
+  },
+  {
+    "id": "us.stability.stable-conservative-upscale-v1:0",
+    "name": "Stable Image Conservative Upscale",
+    "provider": "bedrock",
+    "family": "Stable Image Services",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "image"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Stability AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/stability.stable-conservative-upscale-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "us.stability.stable-creative-upscale-v1:0",
+    "name": "Stable Image Creative Upscale",
+    "provider": "bedrock",
+    "family": "Stable Image Services",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "image"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Stability AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/stability.stable-creative-upscale-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "us.stability.stable-fast-upscale-v1:0",
+    "name": "Stable Image Fast Upscale",
+    "provider": "bedrock",
+    "family": "Stable Image Services",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "image"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Stability AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/stability.stable-fast-upscale-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "us.stability.stable-image-control-sketch-v1:0",
+    "name": "Stable Image Control Sketch",
+    "provider": "bedrock",
+    "family": "Stable Image Services",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "image"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Stability AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/stability.stable-image-control-sketch-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "us.stability.stable-image-control-structure-v1:0",
+    "name": "Stable Image Control Structure",
+    "provider": "bedrock",
+    "family": "Stable Image Services",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "image"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Stability AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/stability.stable-image-control-structure-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "us.stability.stable-image-erase-object-v1:0",
+    "name": "Stable Image Erase Object",
+    "provider": "bedrock",
+    "family": "Stable Image Services",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "image"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Stability AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/stability.stable-image-erase-object-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "us.stability.stable-image-inpaint-v1:0",
+    "name": "Stable Image Inpaint",
+    "provider": "bedrock",
+    "family": "Stable Image Services",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "image"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Stability AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/stability.stable-image-inpaint-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "us.stability.stable-image-remove-background-v1:0",
+    "name": "Stable Image Remove Background",
+    "provider": "bedrock",
+    "family": "Stable Image Services",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "image"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Stability AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/stability.stable-image-remove-background-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "us.stability.stable-image-search-recolor-v1:0",
+    "name": "Stable Image Search and Recolor",
+    "provider": "bedrock",
+    "family": "Stable Image Services",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "image"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Stability AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/stability.stable-image-search-recolor-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "us.stability.stable-image-search-replace-v1:0",
+    "name": "Stable Image Search and Replace",
+    "provider": "bedrock",
+    "family": "Stable Image Services",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "image"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Stability AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/stability.stable-image-search-replace-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "us.stability.stable-image-style-guide-v1:0",
+    "name": "Stable Image Style Guide",
+    "provider": "bedrock",
+    "family": "Stable Image Services",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "image"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Stability AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/stability.stable-image-style-guide-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "us.stability.stable-outpaint-v1:0",
+    "name": "Stable Image Outpaint",
+    "provider": "bedrock",
+    "family": "Stable Image Services",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "image"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Stability AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/stability.stable-outpaint-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "us.stability.stable-style-transfer-v1:0",
+    "name": "Stable Image Style Transfer",
+    "provider": "bedrock",
+    "family": "Stable Image Services",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "image"
+      ]
+    },
+    "capabilities": [
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Stability AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/stability.stable-style-transfer-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "us.twelvelabs.pegasus-1-2-v1:0",
+    "name": "Pegasus v1.2",
+    "provider": "bedrock",
+    "family": "Pegasus",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "TwelveLabs",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/twelvelabs.pegasus-1-2-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {}
+    }
+  },
+  {
+    "id": "us.writer.palmyra-x4-v1:0",
+    "name": "Palmyra X4",
+    "provider": "bedrock",
+    "family": "palmyra",
+    "created_at": "2025-04-28 00:00:00 UTC",
+    "context_window": 122880,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Writer",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/writer.palmyra-x4-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2025-04-28",
+      "cost": {
+        "input": 2.5,
+        "output": 10
+      },
+      "limit": {
+        "context": 122880,
+        "output": 8192
+      }
+    }
+  },
+  {
+    "id": "us.writer.palmyra-x5-v1:0",
+    "name": "Palmyra X5",
+    "provider": "bedrock",
+    "family": "palmyra",
+    "created_at": "2025-04-28 00:00:00 UTC",
+    "context_window": 1040000,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.6,
+          "output_per_million": 6
+        }
+      }
+    },
+    "metadata": {
+      "provider_name": "Writer",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/writer.palmyra-x5-v1:0",
+      "inference_types": [
+        "INFERENCE_PROFILE"
+      ],
+      "converse": {},
+      "source": "models.dev",
+      "provider_id": "amazon-bedrock",
+      "open_weights": false,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2025-04-28",
+      "cost": {
+        "input": 0.6,
+        "output": 6
+      },
+      "limit": {
+        "context": 1040000,
+        "output": 8192
+      }
+    }
+  },
+  {
+    "id": "writer.palmyra-vision-7b",
+    "name": "Writer Palmyra Vision 7B",
+    "provider": "bedrock",
+    "family": "Writer Palmyra Vision",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "provider_name": "Writer",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/writer.palmyra-vision-7b",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 4096,
+        "reasoningSupported": null,
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [
+          "jpeg",
+          "png",
+          "gif",
+          "webp"
+        ],
+        "userVideoTypesSupported": []
+      }
     }
   },
   {
@@ -5280,7 +18194,7 @@
     "name": "Palmyra X4",
     "provider": "bedrock",
     "family": "palmyra",
-    "created_at": "2025-04-28 00:00:00 +0530",
+    "created_at": "2025-04-28 00:00:00 UTC",
     "context_window": 122880,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -5326,7 +18240,7 @@
     "name": "Palmyra X5",
     "provider": "bedrock",
     "family": "palmyra",
-    "created_at": "2025-04-28 00:00:00 +0530",
+    "created_at": "2025-04-28 00:00:00 UTC",
     "context_window": 1040000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -5372,7 +18286,7 @@
     "name": "GLM-4.7",
     "provider": "bedrock",
     "family": "glm",
-    "created_at": "2025-12-22 00:00:00 +0530",
+    "created_at": "2025-12-22 00:00:00 UTC",
     "context_window": 204800,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -5386,7 +18300,8 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -5397,6 +18312,24 @@
       }
     },
     "metadata": {
+      "provider_name": "Z.AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/zai.glm-4.7",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 202752,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -5422,7 +18355,7 @@
     "name": "GLM-4.7-Flash",
     "provider": "bedrock",
     "family": "glm-flash",
-    "created_at": "2026-01-19 00:00:00 +0530",
+    "created_at": "2026-01-19 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -5436,7 +18369,8 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -5447,6 +18381,24 @@
       }
     },
     "metadata": {
+      "provider_name": "Z.AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/zai.glm-4.7-flash",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 202752,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -5469,7 +18421,7 @@
     "name": "GLM-5",
     "provider": "bedrock",
     "family": "glm",
-    "created_at": "2026-03-18 00:00:00 +0530",
+    "created_at": "2026-03-18 00:00:00 UTC",
     "context_window": 202752,
     "max_output_tokens": 101376,
     "knowledge_cutoff": null,
@@ -5483,7 +18435,8 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -5494,6 +18447,24 @@
       }
     },
     "metadata": {
+      "provider_name": "Z.AI",
+      "model_arn": "arn:aws:bedrock:us-west-2::foundation-model/zai.glm-5",
+      "inference_types": [
+        "ON_DEMAND"
+      ],
+      "converse": {
+        "additionalRequestFieldsSchema": "{\"reasoningConfig\":{\"enabled\":{\"type\":\"boolean\",\"default\":false}}}",
+        "maxTokensDefault": null,
+        "maxTokensMaximum": 202752,
+        "reasoningSupported": {
+          "embedded": false
+        },
+        "stopSequencesDefault": [],
+        "systemRoleSupported": true,
+        "userDocumentTypesSupported": [],
+        "userImageTypesSupported": [],
+        "userVideoTypesSupported": []
+      },
       "source": "models.dev",
       "provider_id": "amazon-bedrock",
       "open_weights": true,
@@ -5518,9 +18489,9 @@
     "name": "DeepSeek Chat",
     "provider": "deepseek",
     "family": "deepseek",
-    "created_at": "2025-12-01 00:00:00 +0530",
-    "context_window": 131072,
-    "max_output_tokens": 8192,
+    "created_at": "2025-12-01 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 384000,
     "knowledge_cutoff": null,
     "modalities": {
       "input": [
@@ -5536,9 +18507,9 @@
     "pricing": {
       "text_tokens": {
         "standard": {
-          "input_per_million": 0.28,
-          "output_per_million": 0.42,
-          "cached_input_per_million": 0.028
+          "input_per_million": 0.14,
+          "output_per_million": 0.28,
+          "cache_read_input_per_million": 0.028
         }
       }
     },
@@ -5550,13 +18521,13 @@
       "temperature": true,
       "last_updated": "2026-02-28",
       "cost": {
-        "input": 0.28,
-        "output": 0.42,
+        "input": 0.14,
+        "output": 0.28,
         "cache_read": 0.028
       },
       "limit": {
-        "context": 131072,
-        "output": 8192
+        "context": 1000000,
+        "output": 384000
       },
       "knowledge": "2025-09"
     }
@@ -5566,9 +18537,9 @@
     "name": "DeepSeek Reasoner",
     "provider": "deepseek",
     "family": "deepseek-thinking",
-    "created_at": "2025-12-01 00:00:00 +0530",
-    "context_window": 128000,
-    "max_output_tokens": 64000,
+    "created_at": "2025-12-01 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 384000,
     "knowledge_cutoff": null,
     "modalities": {
       "input": [
@@ -5585,9 +18556,9 @@
     "pricing": {
       "text_tokens": {
         "standard": {
-          "input_per_million": 0.28,
-          "output_per_million": 0.42,
-          "cached_input_per_million": 0.028
+          "input_per_million": 0.14,
+          "output_per_million": 0.28,
+          "cache_read_input_per_million": 0.028
         }
       }
     },
@@ -5602,23 +18573,249 @@
         "field": "reasoning_content"
       },
       "cost": {
-        "input": 0.28,
-        "output": 0.42,
+        "input": 0.14,
+        "output": 0.28,
         "cache_read": 0.028
       },
       "limit": {
-        "context": 128000,
-        "output": 64000
+        "context": 1000000,
+        "output": 384000
       },
       "knowledge": "2025-09"
     }
   },
+  {
+    "id": "deepseek-v4-flash",
+    "name": "DeepSeek V4 Flash",
+    "provider": "deepseek",
+    "family": "deepseek-flash",
+    "created_at": "2026-04-24 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 384000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.14,
+          "output_per_million": 0.28,
+          "cache_read_input_per_million": 0.028
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "deepseek",
+      "open_weights": true,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2026-04-24",
+      "interleaved": {
+        "field": "reasoning_content"
+      },
+      "cost": {
+        "input": 0.14,
+        "output": 0.28,
+        "cache_read": 0.028
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 384000
+      },
+      "knowledge": "2025-05"
+    }
+  },
+  {
+    "id": "deepseek-v4-pro",
+    "name": "DeepSeek V4 Pro",
+    "provider": "deepseek",
+    "family": "deepseek-thinking",
+    "created_at": "2026-04-24 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 384000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.74,
+          "output_per_million": 3.48,
+          "cache_read_input_per_million": 0.145
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "deepseek",
+      "open_weights": true,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2026-04-24",
+      "interleaved": {
+        "field": "reasoning_content"
+      },
+      "cost": {
+        "input": 1.74,
+        "output": 3.48,
+        "cache_read": 0.145
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 384000
+      },
+      "knowledge": "2025-05"
+    }
+  },
+  {
+    "id": "aqa",
+    "name": "Model that performs Attributed Question Answering.",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 7168,
+    "max_output_tokens": 1024,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {},
+    "metadata": {
+      "version": "001",
+      "description": "Model trained to return answers to questions that are grounded in provided sources, along with estimating answerable probability.",
+      "supported_generation_methods": [
+        "generateAnswer"
+      ]
+    }
+  },
+  {
+    "id": "deep-research-max-preview-04-2026",
+    "name": "Deep Research Max Preview (Apr-21-2026)",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 131072,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "deepthink-exp-05-20",
+      "description": "Preview release (April 21st, 2026) of Deep Research Max",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens"
+      ]
+    }
+  },
+  {
+    "id": "deep-research-preview-04-2026",
+    "name": "Deep Research Preview (Apr-21-2026)",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 131072,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "deepthink-exp-05-20",
+      "description": "Preview release (April 21th, 2026) of Deep Research",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens"
+      ]
+    }
+  },
+  {
+    "id": "deep-research-pro-preview-12-2025",
+    "name": "Deep Research Pro Preview (Dec-12-2025)",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 131072,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "deepthink-exp-05-20",
+      "description": "Preview release (December 12th, 2025) of Deep Research Pro",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens"
+      ]
+    }
+  },
   {
     "id": "gemini-1.5-flash",
     "name": "Gemini 1.5 Flash",
     "provider": "gemini",
     "family": "gemini-flash",
-    "created_at": "2024-05-14 00:00:00 +0530",
+    "created_at": "2024-05-14 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -5642,7 +18839,7 @@
         "standard": {
           "input_per_million": 0.075,
           "output_per_million": 0.3,
-          "cached_input_per_million": 0.01875
+          "cache_read_input_per_million": 0.01875
         }
       }
     },
@@ -5670,7 +18867,7 @@
     "name": "Gemini 1.5 Flash-8B",
     "provider": "gemini",
     "family": "gemini-flash",
-    "created_at": "2024-10-03 00:00:00 +0530",
+    "created_at": "2024-10-03 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -5694,7 +18891,7 @@
         "standard": {
           "input_per_million": 0.0375,
           "output_per_million": 0.15,
-          "cached_input_per_million": 0.01
+          "cache_read_input_per_million": 0.01
         }
       }
     },
@@ -5722,7 +18919,7 @@
     "name": "Gemini 1.5 Pro",
     "provider": "gemini",
     "family": "gemini-pro",
-    "created_at": "2024-02-15 00:00:00 +0530",
+    "created_at": "2024-02-15 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -5746,7 +18943,7 @@
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 5,
-          "cached_input_per_million": 0.3125
+          "cache_read_input_per_million": 0.3125
         }
       }
     },
@@ -5774,7 +18971,7 @@
     "name": "Gemini 2.0 Flash",
     "provider": "gemini",
     "family": "gemini-flash",
-    "created_at": "2024-12-11 00:00:00 +0530",
+    "created_at": "2024-12-11 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -5800,11 +18997,19 @@
         "standard": {
           "input_per_million": 0.1,
           "output_per_million": 0.4,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       }
     },
     "metadata": {
+      "version": "2.0",
+      "description": "Gemini 2.0 Flash",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "createCachedContent",
+        "batchGenerateContent"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": false,
@@ -5823,12 +19028,49 @@
       "knowledge": "2024-06"
     }
   },
+  {
+    "id": "gemini-2.0-flash-001",
+    "name": "Gemini 2.0 Flash 001",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 1048576,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.1,
+          "output_per_million": 0.4
+        }
+      }
+    },
+    "metadata": {
+      "version": "2.0",
+      "description": "Stable version of Gemini 2.0 Flash, our fast and versatile multimodal model for scaling across diverse tasks, released in January of 2025.",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "createCachedContent",
+        "batchGenerateContent"
+      ]
+    }
+  },
   {
     "id": "gemini-2.0-flash-lite",
     "name": "Gemini 2.0 Flash Lite",
     "provider": "gemini",
     "family": "gemini-flash-lite",
-    "created_at": "2024-12-11 00:00:00 +0530",
+    "created_at": "2024-12-11 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -5858,6 +19100,14 @@
       }
     },
     "metadata": {
+      "version": "2.0",
+      "description": "Gemini 2.0 Flash-Lite",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "createCachedContent",
+        "batchGenerateContent"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": false,
@@ -5875,12 +19125,82 @@
       "knowledge": "2024-06"
     }
   },
+  {
+    "id": "gemini-2.0-flash-lite-001",
+    "name": "Gemini 2.0 Flash-Lite 001",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 1048576,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "2.0",
+      "description": "Stable version of Gemini 2.0 Flash-Lite",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "createCachedContent",
+        "batchGenerateContent"
+      ]
+    }
+  },
+  {
+    "id": "gemini-2.5-computer-use-preview-10-2025",
+    "name": "Gemini 2.5 Computer Use Preview 10-2025",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 131072,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "Gemini 2.5 Computer Use Preview 10-2025",
+      "description": "Gemini 2.5 Computer Use Preview 10-2025",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens"
+      ]
+    }
+  },
   {
     "id": "gemini-2.5-flash",
     "name": "Gemini 2.5 Flash",
     "provider": "gemini",
     "family": "gemini-flash",
-    "created_at": "2025-03-20 00:00:00 +0530",
+    "created_at": "2025-03-20 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -5907,7 +19227,7 @@
         "standard": {
           "input_per_million": 0.3,
           "output_per_million": 2.5,
-          "cached_input_per_million": 0.075
+          "cache_read_input_per_million": 0.03
         }
       },
       "audio_tokens": {
@@ -5917,6 +19237,14 @@
       }
     },
     "metadata": {
+      "version": "001",
+      "description": "Stable version of Gemini 2.5 Flash, our mid-size multimodal model that supports up to 1 million tokens, released in June of 2025.",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "createCachedContent",
+        "batchGenerateContent"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": false,
@@ -5926,7 +19254,7 @@
       "cost": {
         "input": 0.3,
         "output": 2.5,
-        "cache_read": 0.075,
+        "cache_read": 0.03,
         "input_audio": 1
       },
       "limit": {
@@ -5941,7 +19269,7 @@
     "name": "Gemini 2.5 Flash Image",
     "provider": "gemini",
     "family": "gemini-flash",
-    "created_at": "2025-08-26 00:00:00 +0530",
+    "created_at": "2025-08-26 00:00:00 UTC",
     "context_window": 32768,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -5957,18 +19285,27 @@
     },
     "capabilities": [
       "reasoning",
-      "vision"
+      "vision",
+      "function_calling",
+      "structured_output"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.3,
           "output_per_million": 30,
-          "cached_input_per_million": 0.075
+          "cache_read_input_per_million": 0.075
         }
       }
     },
     "metadata": {
+      "version": "2.0",
+      "description": "Gemini 2.5 Flash Preview Image",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "batchGenerateContent"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": false,
@@ -5992,7 +19329,7 @@
     "name": "Gemini 2.5 Flash Image (Preview)",
     "provider": "gemini",
     "family": "gemini-flash",
-    "created_at": "2025-08-26 00:00:00 +0530",
+    "created_at": "2025-08-26 00:00:00 UTC",
     "context_window": 32768,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -6015,7 +19352,7 @@
         "standard": {
           "input_per_million": 0.3,
           "output_per_million": 30,
-          "cached_input_per_million": 0.075
+          "cache_read_input_per_million": 0.075
         }
       }
     },
@@ -6043,7 +19380,7 @@
     "name": "Gemini 2.5 Flash Lite",
     "provider": "gemini",
     "family": "gemini-flash-lite",
-    "created_at": "2025-06-17 00:00:00 +0530",
+    "created_at": "2025-06-17 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -6070,11 +19407,19 @@
         "standard": {
           "input_per_million": 0.1,
           "output_per_million": 0.4,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       }
     },
     "metadata": {
+      "version": "001",
+      "description": "Stable version of Gemini 2.5 Flash-Lite, released in July of 2025",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "createCachedContent",
+        "batchGenerateContent"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": false,
@@ -6098,7 +19443,7 @@
     "name": "Gemini 2.5 Flash Lite Preview 06-17",
     "provider": "gemini",
     "family": "gemini-flash-lite",
-    "created_at": "2025-06-17 00:00:00 +0530",
+    "created_at": "2025-06-17 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -6124,7 +19469,7 @@
         "standard": {
           "input_per_million": 0.1,
           "output_per_million": 0.4,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       },
       "audio_tokens": {
@@ -6158,7 +19503,7 @@
     "name": "Gemini 2.5 Flash Lite Preview 09-25",
     "provider": "gemini",
     "family": "gemini-flash-lite",
-    "created_at": "2025-09-25 00:00:00 +0530",
+    "created_at": "2025-09-25 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -6185,7 +19530,7 @@
         "standard": {
           "input_per_million": 0.1,
           "output_per_million": 0.4,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       }
     },
@@ -6208,12 +19553,117 @@
       "knowledge": "2025-01"
     }
   },
+  {
+    "id": "gemini-2.5-flash-native-audio-latest",
+    "name": "Gemini 2.5 Flash Native Audio Latest",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 131072,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "Gemini 2.5 Flash Native Audio Latest",
+      "description": "Latest release of Gemini 2.5 Flash Native Audio",
+      "supported_generation_methods": [
+        "countTokens",
+        "bidiGenerateContent"
+      ]
+    }
+  },
+  {
+    "id": "gemini-2.5-flash-native-audio-preview-09-2025",
+    "name": "Gemini 2.5 Flash Native Audio Preview 09-2025",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 131072,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "gemini-2.5-flash-preview-native-audio-dialog-2025-05-19",
+      "description": "Gemini 2.5 Flash Native Audio Preview 09-2025",
+      "supported_generation_methods": [
+        "countTokens",
+        "bidiGenerateContent"
+      ]
+    }
+  },
+  {
+    "id": "gemini-2.5-flash-native-audio-preview-12-2025",
+    "name": "Gemini 2.5 Flash Native Audio Preview 12-2025",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 131072,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "12-2025",
+      "description": "Gemini 2.5 Flash Native Audio Preview 12-2025",
+      "supported_generation_methods": [
+        "countTokens",
+        "bidiGenerateContent"
+      ]
+    }
+  },
   {
     "id": "gemini-2.5-flash-preview-04-17",
     "name": "Gemini 2.5 Flash Preview 04-17",
     "provider": "gemini",
     "family": "gemini-flash",
-    "created_at": "2025-04-17 00:00:00 +0530",
+    "created_at": "2025-04-17 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -6239,7 +19689,7 @@
         "standard": {
           "input_per_million": 0.15,
           "output_per_million": 0.6,
-          "cached_input_per_million": 0.0375
+          "cache_read_input_per_million": 0.0375
         }
       }
     },
@@ -6267,7 +19717,7 @@
     "name": "Gemini 2.5 Flash Preview 05-20",
     "provider": "gemini",
     "family": "gemini-flash",
-    "created_at": "2025-05-20 00:00:00 +0530",
+    "created_at": "2025-05-20 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -6294,7 +19744,7 @@
         "standard": {
           "input_per_million": 0.15,
           "output_per_million": 0.6,
-          "cached_input_per_million": 0.0375
+          "cache_read_input_per_million": 0.0375
         }
       }
     },
@@ -6322,7 +19772,7 @@
     "name": "Gemini 2.5 Flash Preview 09-25",
     "provider": "gemini",
     "family": "gemini-flash",
-    "created_at": "2025-09-25 00:00:00 +0530",
+    "created_at": "2025-09-25 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -6349,7 +19799,7 @@
         "standard": {
           "input_per_million": 0.3,
           "output_per_million": 2.5,
-          "cached_input_per_million": 0.075
+          "cache_read_input_per_million": 0.075
         }
       },
       "audio_tokens": {
@@ -6383,7 +19833,7 @@
     "name": "Gemini 2.5 Flash Preview TTS",
     "provider": "gemini",
     "family": "gemini-flash",
-    "created_at": "2025-05-01 00:00:00 +0530",
+    "created_at": "2025-05-01 00:00:00 UTC",
     "context_window": 8000,
     "max_output_tokens": 16000,
     "knowledge_cutoff": null,
@@ -6395,7 +19845,11 @@
         "audio"
       ]
     },
-    "capabilities": [],
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -6405,6 +19859,12 @@
       }
     },
     "metadata": {
+      "version": "gemini-2.5-flash-exp-tts-2025-05-19",
+      "description": "Gemini 2.5 Flash Preview TTS",
+      "supported_generation_methods": [
+        "countTokens",
+        "generateContent"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": false,
@@ -6427,7 +19887,7 @@
     "name": "Gemini 2.5 Pro",
     "provider": "gemini",
     "family": "gemini-pro",
-    "created_at": "2025-03-20 00:00:00 +0530",
+    "created_at": "2025-03-20 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -6454,11 +19914,19 @@
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.31
+          "cache_read_input_per_million": 0.125
         }
       }
     },
     "metadata": {
+      "version": "2.5",
+      "description": "Stable release (June 17th, 2025) of Gemini 2.5 Pro",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "createCachedContent",
+        "batchGenerateContent"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": false,
@@ -6468,7 +19936,12 @@
       "cost": {
         "input": 1.25,
         "output": 10,
-        "cache_read": 0.31
+        "cache_read": 0.125,
+        "context_over_200k": {
+          "input": 2.5,
+          "output": 15,
+          "cache_read": 0.25
+        }
       },
       "limit": {
         "context": 1048576,
@@ -6482,7 +19955,7 @@
     "name": "Gemini 2.5 Pro Preview 05-06",
     "provider": "gemini",
     "family": "gemini-pro",
-    "created_at": "2025-05-06 00:00:00 +0530",
+    "created_at": "2025-05-06 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -6509,7 +19982,7 @@
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.31
+          "cache_read_input_per_million": 0.31
         }
       }
     },
@@ -6537,7 +20010,7 @@
     "name": "Gemini 2.5 Pro Preview 06-05",
     "provider": "gemini",
     "family": "gemini-pro",
-    "created_at": "2025-06-05 00:00:00 +0530",
+    "created_at": "2025-06-05 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -6564,7 +20037,7 @@
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.31
+          "cache_read_input_per_million": 0.31
         }
       }
     },
@@ -6592,7 +20065,7 @@
     "name": "Gemini 2.5 Pro Preview TTS",
     "provider": "gemini",
     "family": "gemini-flash",
-    "created_at": "2025-05-01 00:00:00 +0530",
+    "created_at": "2025-05-01 00:00:00 UTC",
     "context_window": 8000,
     "max_output_tokens": 16000,
     "knowledge_cutoff": null,
@@ -6604,7 +20077,11 @@
         "audio"
       ]
     },
-    "capabilities": [],
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -6614,6 +20091,13 @@
       }
     },
     "metadata": {
+      "version": "gemini-2.5-pro-preview-tts-2025-05-19",
+      "description": "Gemini 2.5 Pro Preview TTS",
+      "supported_generation_methods": [
+        "countTokens",
+        "generateContent",
+        "batchGenerateContent"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": false,
@@ -6636,7 +20120,7 @@
     "name": "Gemini 3 Flash Preview",
     "provider": "gemini",
     "family": "gemini-flash",
-    "created_at": "2025-12-17 00:00:00 +0530",
+    "created_at": "2025-12-17 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -6663,11 +20147,19 @@
         "standard": {
           "input_per_million": 0.5,
           "output_per_million": 3,
-          "cached_input_per_million": 0.05
+          "cache_read_input_per_million": 0.05
         }
       }
     },
     "metadata": {
+      "version": "3-flash-preview-12-2025",
+      "description": "Gemini 3 Flash Preview",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "createCachedContent",
+        "batchGenerateContent"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": false,
@@ -6691,12 +20183,48 @@
       "knowledge": "2025-01"
     }
   },
+  {
+    "id": "gemini-3-pro-image-preview",
+    "name": "Nano Banana Pro",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 131072,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "3.0",
+      "description": "Gemini 3 Pro Image Preview",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "batchGenerateContent"
+      ]
+    }
+  },
   {
     "id": "gemini-3-pro-preview",
     "name": "Gemini 3 Pro Preview",
     "provider": "gemini",
     "family": "gemini-pro",
-    "created_at": "2025-11-18 00:00:00 +0530",
+    "created_at": "2025-11-18 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": null,
@@ -6723,11 +20251,19 @@
         "standard": {
           "input_per_million": 2,
           "output_per_million": 12,
-          "cached_input_per_million": 0.2
+          "cache_read_input_per_million": 0.2
         }
       }
     },
     "metadata": {
+      "version": "3-pro-preview-11-2025",
+      "description": "Gemini 3 Pro Preview",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "createCachedContent",
+        "batchGenerateContent"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": false,
@@ -6756,7 +20292,7 @@
     "name": "Gemini 3.1 Flash Image (Preview)",
     "provider": "gemini",
     "family": "gemini-flash",
-    "created_at": "2026-02-26 00:00:00 +0530",
+    "created_at": "2026-02-26 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -6773,7 +20309,9 @@
     },
     "capabilities": [
       "reasoning",
-      "vision"
+      "vision",
+      "function_calling",
+      "structured_output"
     ],
     "pricing": {
       "text_tokens": {
@@ -6784,6 +20322,13 @@
       }
     },
     "metadata": {
+      "version": "3.0",
+      "description": "Gemini 3.1 Flash Image Preview.",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "batchGenerateContent"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": false,
@@ -6806,7 +20351,7 @@
     "name": "Gemini 3.1 Flash Lite Preview",
     "provider": "gemini",
     "family": "gemini-flash-lite",
-    "created_at": "2026-03-03 00:00:00 +0530",
+    "created_at": "2026-03-03 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -6833,11 +20378,20 @@
         "standard": {
           "input_per_million": 0.25,
           "output_per_million": 1.5,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025,
+          "cache_write_input_per_million": 1
         }
       }
     },
     "metadata": {
+      "version": "3.1-flash-lite-preview-03-2026",
+      "description": "Gemini 3.1 Flash Lite Preview",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "createCachedContent",
+        "batchGenerateContent"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": false,
@@ -6857,12 +20411,82 @@
       "knowledge": "2025-01"
     }
   },
+  {
+    "id": "gemini-3.1-flash-live-preview",
+    "name": "Gemini 3.1 Flash Live Preview",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 131072,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "3.1-flash-live-03-2026",
+      "description": "Gemini 3.1 Flash Live Preview",
+      "supported_generation_methods": [
+        "bidiGenerateContent"
+      ]
+    }
+  },
+  {
+    "id": "gemini-3.1-flash-tts-preview",
+    "name": "Gemini 3.1 Flash TTS Preview",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 8192,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "3.1-flash-tts-preview",
+      "description": "Gemini 3.1 Flash TTS Preview",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "batchGenerateContent"
+      ]
+    }
+  },
   {
     "id": "gemini-3.1-pro-preview",
     "name": "Gemini 3.1 Pro Preview",
     "provider": "gemini",
     "family": "gemini-pro",
-    "created_at": "2026-02-19 00:00:00 +0530",
+    "created_at": "2026-02-19 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -6889,11 +20513,19 @@
         "standard": {
           "input_per_million": 2,
           "output_per_million": 12,
-          "cached_input_per_million": 0.2
+          "cache_read_input_per_million": 0.2
         }
       }
     },
     "metadata": {
+      "version": "3.1-pro-preview-01-2026",
+      "description": "Gemini 3.1 Pro Preview",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "createCachedContent",
+        "batchGenerateContent"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": false,
@@ -6922,7 +20554,7 @@
     "name": "Gemini 3.1 Pro Preview Custom Tools",
     "provider": "gemini",
     "family": "gemini-pro",
-    "created_at": "2026-02-19 00:00:00 +0530",
+    "created_at": "2026-02-19 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -6949,11 +20581,19 @@
         "standard": {
           "input_per_million": 2,
           "output_per_million": 12,
-          "cached_input_per_million": 0.2
+          "cache_read_input_per_million": 0.2
         }
       }
     },
     "metadata": {
+      "version": "3.1-pro-preview-01-2026",
+      "description": "Gemini 3.1 Pro Preview optimized for custom tool usage",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "createCachedContent",
+        "batchGenerateContent"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": false,
@@ -6982,7 +20622,7 @@
     "name": "Gemini Embedding 001",
     "provider": "gemini",
     "family": "gemini",
-    "created_at": "2025-05-20 00:00:00 +0530",
+    "created_at": "2025-05-20 00:00:00 UTC",
     "context_window": 2048,
     "max_output_tokens": 3072,
     "knowledge_cutoff": null,
@@ -6991,7 +20631,7 @@
         "text"
       ],
       "output": [
-        "text"
+        "embeddings"
       ]
     },
     "capabilities": [],
@@ -7003,6 +20643,14 @@
       }
     },
     "metadata": {
+      "version": "001",
+      "description": "Obtain a distributed representation of a text.",
+      "supported_generation_methods": [
+        "embedContent",
+        "countTextTokens",
+        "countTokens",
+        "asyncBatchEmbedContent"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": false,
@@ -7020,12 +20668,72 @@
       "knowledge": "2025-05"
     }
   },
+  {
+    "id": "gemini-embedding-2",
+    "name": "Gemini Embedding 2",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 8192,
+    "max_output_tokens": 1,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {},
+    "metadata": {
+      "version": "2",
+      "description": "Obtain a distributed representation of multimodal content.",
+      "supported_generation_methods": [
+        "embedContent",
+        "countTextTokens",
+        "countTokens",
+        "asyncBatchEmbedContent"
+      ]
+    }
+  },
+  {
+    "id": "gemini-embedding-2-preview",
+    "name": "Gemini Embedding 2 Preview",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 8192,
+    "max_output_tokens": 1,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {},
+    "metadata": {
+      "version": "2",
+      "description": "Obtain a distributed representation of multimodal content.",
+      "supported_generation_methods": [
+        "embedContent",
+        "countTextTokens",
+        "countTokens",
+        "asyncBatchEmbedContent"
+      ]
+    }
+  },
   {
     "id": "gemini-flash-latest",
     "name": "Gemini Flash Latest",
     "provider": "gemini",
     "family": "gemini-flash",
-    "created_at": "2025-09-25 00:00:00 +0530",
+    "created_at": "2025-09-25 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -7052,7 +20760,7 @@
         "standard": {
           "input_per_million": 0.3,
           "output_per_million": 2.5,
-          "cached_input_per_million": 0.075
+          "cache_read_input_per_million": 0.075
         }
       },
       "audio_tokens": {
@@ -7062,6 +20770,14 @@
       }
     },
     "metadata": {
+      "version": "Gemini Flash Latest",
+      "description": "Latest release of Gemini Flash",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "createCachedContent",
+        "batchGenerateContent"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": false,
@@ -7086,7 +20802,7 @@
     "name": "Gemini Flash-Lite Latest",
     "provider": "gemini",
     "family": "gemini-flash-lite",
-    "created_at": "2025-09-25 00:00:00 +0530",
+    "created_at": "2025-09-25 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -7113,11 +20829,19 @@
         "standard": {
           "input_per_million": 0.1,
           "output_per_million": 0.4,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       }
     },
     "metadata": {
+      "version": "Gemini Flash-Lite Latest",
+      "description": "Latest release of Gemini Flash-Lite",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "createCachedContent",
+        "batchGenerateContent"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": false,
@@ -7141,7 +20865,7 @@
     "name": "Gemini Live 2.5 Flash",
     "provider": "gemini",
     "family": "gemini-flash",
-    "created_at": "2025-09-01 00:00:00 +0530",
+    "created_at": "2025-09-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 8000,
     "knowledge_cutoff": null,
@@ -7201,7 +20925,7 @@
     "name": "Gemini Live 2.5 Flash Preview Native Audio",
     "provider": "gemini",
     "family": "gemini-flash",
-    "created_at": "2025-06-17 00:00:00 +0530",
+    "created_at": "2025-06-17 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -7255,12 +20979,121 @@
       "knowledge": "2025-01"
     }
   },
+  {
+    "id": "gemini-pro-latest",
+    "name": "Gemini Pro Latest",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 1048576,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "Gemini Pro Latest",
+      "description": "Latest release of Gemini Pro",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "createCachedContent",
+        "batchGenerateContent"
+      ]
+    }
+  },
+  {
+    "id": "gemini-robotics-er-1.5-preview",
+    "name": "Gemini Robotics-ER 1.5 Preview",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 1048576,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "1.5-preview",
+      "description": "Gemini Robotics-ER 1.5 Preview",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens"
+      ]
+    }
+  },
+  {
+    "id": "gemini-robotics-er-1.6-preview",
+    "name": "Gemini Robotics-ER 1.6 Preview",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 131072,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "1.6-preview",
+      "description": "Gemini Robotics-ER 1.6 Preview",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "createCachedContent",
+        "batchGenerateContent"
+      ]
+    }
+  },
   {
     "id": "gemma-3-12b-it",
     "name": "Gemma 3 12B",
     "provider": "gemini",
     "family": "gemma",
-    "created_at": "2025-03-13 00:00:00 +0530",
+    "created_at": "2025-03-13 00:00:00 UTC",
     "context_window": 32768,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -7301,7 +21134,7 @@
     "name": "Gemma 3 27B",
     "provider": "gemini",
     "family": "gemma",
-    "created_at": "2025-03-12 00:00:00 +0530",
+    "created_at": "2025-03-12 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -7343,7 +21176,7 @@
     "name": "Gemma 3 4B",
     "provider": "gemini",
     "family": "gemma",
-    "created_at": "2025-03-13 00:00:00 +0530",
+    "created_at": "2025-03-13 00:00:00 UTC",
     "context_window": 32768,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -7383,7 +21216,7 @@
     "name": "Gemma 3n 2B",
     "provider": "gemini",
     "family": "gemma",
-    "created_at": "2025-07-09 00:00:00 +0530",
+    "created_at": "2025-07-09 00:00:00 UTC",
     "context_window": 8192,
     "max_output_tokens": 2000,
     "knowledge_cutoff": null,
@@ -7420,7 +21253,7 @@
     "name": "Gemma 3n 4B",
     "provider": "gemini",
     "family": "gemma",
-    "created_at": "2025-05-20 00:00:00 +0530",
+    "created_at": "2025-05-20 00:00:00 UTC",
     "context_window": 8192,
     "max_output_tokens": 2000,
     "knowledge_cutoff": null,
@@ -7453,11 +21286,11 @@
     }
   },
   {
-    "id": "gemma-4-26b-it",
+    "id": "gemma-4-26b-a4b-it",
     "name": "Gemma 4 26B",
     "provider": "gemini",
     "family": "gemma",
-    "created_at": "2026-04-02 00:00:00 +0530",
+    "created_at": "2026-04-02 00:00:00 UTC",
     "context_window": 256000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -7476,8 +21309,21 @@
       "reasoning",
       "vision"
     ],
-    "pricing": {},
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
     "metadata": {
+      "version": "001",
+      "description": "Gemma 4 26B A4B IT",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": true,
@@ -7495,7 +21341,7 @@
     "name": "Gemma 4 31B",
     "provider": "gemini",
     "family": "gemma",
-    "created_at": "2026-04-02 00:00:00 +0530",
+    "created_at": "2026-04-02 00:00:00 UTC",
     "context_window": 256000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -7514,8 +21360,21 @@
       "reasoning",
       "vision"
     ],
-    "pricing": {},
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
     "metadata": {
+      "version": "001",
+      "description": "Gemma 4 31B IT",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens"
+      ],
       "source": "models.dev",
       "provider_id": "google",
       "open_weights": true,
@@ -7528,12 +21387,472 @@
       }
     }
   },
+  {
+    "id": "imagen-4.0-fast-generate-001",
+    "name": "Imagen 4 Fast",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 480,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.03,
+          "output_per_million": 0.03
+        }
+      }
+    },
+    "metadata": {
+      "version": "001",
+      "description": "Vertex served Imagen 4.0 Fast model",
+      "supported_generation_methods": [
+        "predict"
+      ]
+    }
+  },
+  {
+    "id": "imagen-4.0-generate-001",
+    "name": "Imagen 4",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 480,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.03,
+          "output_per_million": 0.03
+        }
+      }
+    },
+    "metadata": {
+      "version": "001",
+      "description": "Vertex served Imagen 4.0 model",
+      "supported_generation_methods": [
+        "predict"
+      ]
+    }
+  },
+  {
+    "id": "imagen-4.0-ultra-generate-001",
+    "name": "Imagen 4 Ultra",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 480,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.03,
+          "output_per_million": 0.03
+        }
+      }
+    },
+    "metadata": {
+      "version": "001",
+      "description": "Vertex served Imagen 4.0 ultra model",
+      "supported_generation_methods": [
+        "predict"
+      ]
+    }
+  },
+  {
+    "id": "lyria-3-clip-preview",
+    "name": "Lyria 3 Clip Preview",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 1048576,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "lyria-3-clip-preview",
+      "description": "Lyria 3 30s model Preview",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens"
+      ]
+    }
+  },
+  {
+    "id": "lyria-3-pro-preview",
+    "name": "Lyria 3 Pro Preview",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 1048576,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "lyria-3-pro-preview",
+      "description": "Lyria 3 Pro Preview",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens"
+      ]
+    }
+  },
+  {
+    "id": "nano-banana-pro-preview",
+    "name": "Nano Banana Pro",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 131072,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "3.0",
+      "description": "Gemini 3 Pro Image Preview",
+      "supported_generation_methods": [
+        "generateContent",
+        "countTokens",
+        "batchGenerateContent"
+      ]
+    }
+  },
+  {
+    "id": "veo-2.0-generate-001",
+    "name": "Veo 2",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 480,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "2.0",
+      "description": "Vertex served Veo 2 model. Access to this model requires billing to be enabled on the associated Google Cloud Platform account. Please visit https://console.cloud.google.com/billing to enable it.",
+      "supported_generation_methods": [
+        "predictLongRunning"
+      ]
+    }
+  },
+  {
+    "id": "veo-3.0-fast-generate-001",
+    "name": "Veo 3 fast",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 480,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "3.0",
+      "description": "Veo 3 fast",
+      "supported_generation_methods": [
+        "predictLongRunning"
+      ]
+    }
+  },
+  {
+    "id": "veo-3.0-generate-001",
+    "name": "Veo 3",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 480,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "3.0",
+      "description": "Veo 3",
+      "supported_generation_methods": [
+        "predictLongRunning"
+      ]
+    }
+  },
+  {
+    "id": "veo-3.1-fast-generate-preview",
+    "name": "Veo 3.1 fast",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 480,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "3.1",
+      "description": "Veo 3.1 fast",
+      "supported_generation_methods": [
+        "predictLongRunning"
+      ]
+    }
+  },
+  {
+    "id": "veo-3.1-generate-preview",
+    "name": "Veo 3.1",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 480,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "3.1",
+      "description": "Veo 3.1",
+      "supported_generation_methods": [
+        "predictLongRunning"
+      ]
+    }
+  },
+  {
+    "id": "veo-3.1-lite-generate-preview",
+    "name": "Veo 3.1 lite",
+    "provider": "gemini",
+    "family": null,
+    "created_at": null,
+    "context_window": 480,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "version": "3.1",
+      "description": "Veo 3.1 lite",
+      "supported_generation_methods": [
+        "predictLongRunning"
+      ]
+    }
+  },
+  {
+    "id": "codestral-2508",
+    "name": "Codestral",
+    "provider": "mistral",
+    "family": "codestral",
+    "created_at": "2025-08-29 23:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "batch",
+      "predicted_outputs"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "codestral-embed",
+    "name": "Codestral",
+    "provider": "mistral",
+    "family": "codestral",
+    "created_at": "2025-05-20 23:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "embeddings"
+      ]
+    },
+    "capabilities": [
+      "predicted_outputs"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "codestral-embed-2505",
+    "name": "Codestral",
+    "provider": "mistral",
+    "family": "codestral",
+    "created_at": "2025-05-20 23:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "embeddings"
+      ]
+    },
+    "capabilities": [
+      "predicted_outputs"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
   {
     "id": "codestral-latest",
     "name": "Codestral (latest)",
     "provider": "mistral",
     "family": "codestral",
-    "created_at": "2024-05-29 00:00:00 +0530",
+    "created_at": "2024-05-29 00:00:00 UTC",
     "context_window": 256000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -7546,7 +21865,11 @@
       ]
     },
     "capabilities": [
-      "function_calling"
+      "function_calling",
+      "streaming",
+      "structured_output",
+      "batch",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -7557,6 +21880,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": true,
@@ -7579,7 +21904,7 @@
     "name": "Devstral 2",
     "provider": "mistral",
     "family": "devstral",
-    "created_at": "2025-12-09 00:00:00 +0530",
+    "created_at": "2025-12-09 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 262144,
     "knowledge_cutoff": null,
@@ -7592,7 +21917,11 @@
       ]
     },
     "capabilities": [
-      "function_calling"
+      "function_calling",
+      "streaming",
+      "structured_output",
+      "batch",
+      "fine_tuning"
     ],
     "pricing": {
       "text_tokens": {
@@ -7603,6 +21932,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": true,
@@ -7620,12 +21951,42 @@
       "knowledge": "2025-12"
     }
   },
+  {
+    "id": "devstral-latest",
+    "name": "Devstral Latest",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "batch",
+      "fine_tuning"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
   {
     "id": "devstral-medium-2507",
     "name": "Devstral Medium",
     "provider": "mistral",
     "family": "devstral",
-    "created_at": "2025-07-10 00:00:00 +0530",
+    "created_at": "2025-07-10 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": null,
@@ -7638,7 +21999,11 @@
       ]
     },
     "capabilities": [
-      "function_calling"
+      "function_calling",
+      "streaming",
+      "structured_output",
+      "batch",
+      "fine_tuning"
     ],
     "pricing": {
       "text_tokens": {
@@ -7649,6 +22014,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": true,
@@ -7671,7 +22038,7 @@
     "name": "Devstral 2 (latest)",
     "provider": "mistral",
     "family": "devstral",
-    "created_at": "2025-12-02 00:00:00 +0530",
+    "created_at": "2025-12-02 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 262144,
     "knowledge_cutoff": null,
@@ -7684,7 +22051,11 @@
       ]
     },
     "capabilities": [
-      "function_calling"
+      "function_calling",
+      "streaming",
+      "structured_output",
+      "batch",
+      "fine_tuning"
     ],
     "pricing": {
       "text_tokens": {
@@ -7695,6 +22066,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": true,
@@ -7717,7 +22090,7 @@
     "name": "Devstral Small 2505",
     "provider": "mistral",
     "family": "devstral",
-    "created_at": "2025-05-07 00:00:00 +0530",
+    "created_at": "2025-05-07 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": null,
@@ -7763,7 +22136,7 @@
     "name": "Devstral Small",
     "provider": "mistral",
     "family": "devstral",
-    "created_at": "2025-07-10 00:00:00 +0530",
+    "created_at": "2025-07-10 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": null,
@@ -7776,7 +22149,11 @@
       ]
     },
     "capabilities": [
-      "function_calling"
+      "function_calling",
+      "streaming",
+      "structured_output",
+      "batch",
+      "fine_tuning"
     ],
     "pricing": {
       "text_tokens": {
@@ -7787,6 +22164,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": true,
@@ -7809,7 +22188,7 @@
     "name": "Devstral Small 2",
     "provider": "mistral",
     "family": "devstral",
-    "created_at": "2025-12-09 00:00:00 +0530",
+    "created_at": "2025-12-09 00:00:00 UTC",
     "context_window": 256000,
     "max_output_tokens": 256000,
     "knowledge_cutoff": null,
@@ -7845,12 +22224,71 @@
       "knowledge": "2025-12"
     }
   },
+  {
+    "id": "labs-leanstral-2603",
+    "name": "Labs Leanstral 2603",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "batch"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "magistral-medium-2509",
+    "name": "Magistral Medium 2509",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "batch"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
   {
     "id": "magistral-medium-latest",
     "name": "Magistral Medium (latest)",
     "provider": "mistral",
     "family": "magistral-medium",
-    "created_at": "2025-03-17 00:00:00 +0530",
+    "created_at": "2025-03-17 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
@@ -7864,7 +22302,10 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "structured_output",
+      "batch"
     ],
     "pricing": {
       "text_tokens": {
@@ -7875,6 +22316,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": true,
@@ -7897,7 +22340,7 @@
     "name": "Magistral Small",
     "provider": "mistral",
     "family": "magistral-small",
-    "created_at": "2025-03-17 00:00:00 +0530",
+    "created_at": "2025-03-17 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": null,
@@ -7939,12 +22382,162 @@
       "knowledge": "2025-06"
     }
   },
+  {
+    "id": "magistral-small-2509",
+    "name": "Magistral Small 2509",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "batch"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "magistral-small-latest",
+    "name": "Magistral Small Latest",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "batch"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "ministral-14b-2512",
+    "name": "Ministral 14b 2512",
+    "provider": "mistral",
+    "family": "ministral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "batch",
+      "distillation"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "ministral-14b-latest",
+    "name": "Ministral 14b Latest",
+    "provider": "mistral",
+    "family": "ministral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "batch",
+      "distillation"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "ministral-3b-2512",
+    "name": "Ministral 3B",
+    "provider": "mistral",
+    "family": "ministral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "batch",
+      "distillation"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
   {
     "id": "ministral-3b-latest",
     "name": "Ministral 3B (latest)",
     "provider": "mistral",
     "family": "ministral",
-    "created_at": "2024-10-01 00:00:00 +0530",
+    "created_at": "2024-10-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": null,
@@ -7957,7 +22550,11 @@
       ]
     },
     "capabilities": [
-      "function_calling"
+      "function_calling",
+      "streaming",
+      "structured_output",
+      "batch",
+      "distillation"
     ],
     "pricing": {
       "text_tokens": {
@@ -7968,6 +22565,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": true,
@@ -7985,12 +22584,42 @@
       "knowledge": "2024-10"
     }
   },
+  {
+    "id": "ministral-8b-2512",
+    "name": "Ministral 8B",
+    "provider": "mistral",
+    "family": "ministral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "batch",
+      "distillation"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
   {
     "id": "ministral-8b-latest",
     "name": "Ministral 8B (latest)",
     "provider": "mistral",
     "family": "ministral",
-    "created_at": "2024-10-01 00:00:00 +0530",
+    "created_at": "2024-10-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": null,
@@ -8003,7 +22632,11 @@
       ]
     },
     "capabilities": [
-      "function_calling"
+      "function_calling",
+      "streaming",
+      "structured_output",
+      "batch",
+      "distillation"
     ],
     "pricing": {
       "text_tokens": {
@@ -8014,6 +22647,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": true,
@@ -8036,7 +22671,7 @@
     "name": "Mistral Embed",
     "provider": "mistral",
     "family": "mistral-embed",
-    "created_at": "2023-12-11 00:00:00 +0530",
+    "created_at": "2023-12-11 00:00:00 UTC",
     "context_window": 8000,
     "max_output_tokens": 3072,
     "knowledge_cutoff": null,
@@ -8057,6 +22692,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": false,
@@ -8073,12 +22710,36 @@
       }
     }
   },
+  {
+    "id": "mistral-embed-2312",
+    "name": "Mistral Embed",
+    "provider": "mistral",
+    "family": "mistral-embed",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "embeddings"
+      ]
+    },
+    "capabilities": [],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
   {
     "id": "mistral-large-2411",
     "name": "Mistral Large 2.1",
     "provider": "mistral",
     "family": "mistral-large",
-    "created_at": "2024-11-01 00:00:00 +0530",
+    "created_at": "2024-11-01 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
@@ -8091,7 +22752,11 @@
       ]
     },
     "capabilities": [
-      "function_calling"
+      "function_calling",
+      "streaming",
+      "structured_output",
+      "batch",
+      "fine_tuning"
     ],
     "pricing": {
       "text_tokens": {
@@ -8102,6 +22767,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": true,
@@ -8124,7 +22791,7 @@
     "name": "Mistral Large 3",
     "provider": "mistral",
     "family": "mistral-large",
-    "created_at": "2024-11-01 00:00:00 +0530",
+    "created_at": "2024-11-01 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 262144,
     "knowledge_cutoff": null,
@@ -8139,7 +22806,11 @@
     },
     "capabilities": [
       "function_calling",
-      "vision"
+      "vision",
+      "streaming",
+      "structured_output",
+      "batch",
+      "fine_tuning"
     ],
     "pricing": {
       "text_tokens": {
@@ -8150,6 +22821,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": true,
@@ -8172,7 +22845,7 @@
     "name": "Mistral Large (latest)",
     "provider": "mistral",
     "family": "mistral-large",
-    "created_at": "2024-11-01 00:00:00 +0530",
+    "created_at": "2024-11-01 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 262144,
     "knowledge_cutoff": null,
@@ -8187,7 +22860,11 @@
     },
     "capabilities": [
       "function_calling",
-      "vision"
+      "vision",
+      "streaming",
+      "structured_output",
+      "batch",
+      "fine_tuning"
     ],
     "pricing": {
       "text_tokens": {
@@ -8198,6 +22875,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": true,
@@ -8215,12 +22894,75 @@
       "knowledge": "2024-11"
     }
   },
+  {
+    "id": "mistral-large-pixtral-2411",
+    "name": "Mistral Large",
+    "provider": "mistral",
+    "family": "mistral-large",
+    "created_at": "2024-11-12 00:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "vision",
+      "batch",
+      "fine_tuning"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "mistral-medium",
+    "name": "Mistral Medium",
+    "provider": "mistral",
+    "family": "mistral-medium",
+    "created_at": "2025-05-05 23:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "vision",
+      "batch",
+      "fine_tuning"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
   {
     "id": "mistral-medium-2505",
     "name": "Mistral Medium 3",
     "provider": "mistral",
     "family": "mistral-medium",
-    "created_at": "2025-05-07 00:00:00 +0530",
+    "created_at": "2025-05-07 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -8235,7 +22977,11 @@
     },
     "capabilities": [
       "function_calling",
-      "vision"
+      "vision",
+      "streaming",
+      "structured_output",
+      "batch",
+      "fine_tuning"
     ],
     "pricing": {
       "text_tokens": {
@@ -8246,6 +22992,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": false,
@@ -8268,7 +23016,7 @@
     "name": "Mistral Medium 3.1",
     "provider": "mistral",
     "family": "mistral-medium",
-    "created_at": "2025-08-12 00:00:00 +0530",
+    "created_at": "2025-08-12 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 262144,
     "knowledge_cutoff": null,
@@ -8283,7 +23031,11 @@
     },
     "capabilities": [
       "function_calling",
-      "vision"
+      "vision",
+      "streaming",
+      "structured_output",
+      "batch",
+      "fine_tuning"
     ],
     "pricing": {
       "text_tokens": {
@@ -8294,6 +23046,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": false,
@@ -8312,13 +23066,13 @@
     }
   },
   {
-    "id": "mistral-medium-latest",
-    "name": "Mistral Medium (latest)",
+    "id": "mistral-medium-2604",
+    "name": "Mistral Medium 3.5",
     "provider": "mistral",
     "family": "mistral-medium",
-    "created_at": "2025-05-07 00:00:00 +0530",
-    "context_window": 128000,
-    "max_output_tokens": 16384,
+    "created_at": "2026-04-29 00:00:00 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 262144,
     "knowledge_cutoff": null,
     "modalities": {
       "input": [
@@ -8331,32 +23085,294 @@
     },
     "capabilities": [
       "function_calling",
-      "vision"
+      "structured_output",
+      "reasoning",
+      "vision",
+      "streaming",
+      "batch",
+      "fine_tuning"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
-          "input_per_million": 0.4,
-          "output_per_million": 2
+          "input_per_million": 1.5,
+          "output_per_million": 7.5
         }
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": true,
-      "attachment": false,
+      "attachment": true,
       "temperature": true,
-      "last_updated": "2025-05-10",
+      "last_updated": "2026-04-29",
       "cost": {
-        "input": 0.4,
-        "output": 2
+        "input": 1.5,
+        "output": 7.5
       },
       "limit": {
-        "context": 128000,
-        "output": 16384
+        "context": 262144,
+        "output": 262144
+      }
+    }
+  },
+  {
+    "id": "mistral-medium-3",
+    "name": "Mistral Medium",
+    "provider": "mistral",
+    "family": "mistral-medium",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "vision",
+      "batch",
+      "fine_tuning"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "mistral-medium-3-5",
+    "name": "Mistral Medium",
+    "provider": "mistral",
+    "family": "mistral-medium",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "vision",
+      "batch",
+      "fine_tuning"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "mistral-medium-3.5",
+    "name": "Mistral Medium",
+    "provider": "mistral",
+    "family": "mistral-medium",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "vision",
+      "batch",
+      "fine_tuning"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "mistral-medium-c21211-r0-75",
+    "name": "Mistral Medium",
+    "provider": "mistral",
+    "family": "mistral-medium",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "vision",
+      "batch",
+      "fine_tuning"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "mistral-medium-latest",
+    "name": "Mistral Medium (latest)",
+    "provider": "mistral",
+    "family": "mistral-medium",
+    "created_at": "2026-04-29 00:00:00 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 262144,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "vision",
+      "streaming",
+      "batch",
+      "fine_tuning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.5,
+          "output_per_million": 7.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
+      "source": "models.dev",
+      "provider_id": "mistral",
+      "open_weights": true,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2026-04-29",
+      "cost": {
+        "input": 1.5,
+        "output": 7.5
       },
-      "knowledge": "2025-05"
+      "limit": {
+        "context": 262144,
+        "output": 262144
+      }
+    }
+  },
+  {
+    "id": "mistral-moderation-2411",
+    "name": "Mistral Moderation",
+    "provider": "mistral",
+    "family": "mistral-moderation",
+    "created_at": "2024-11-26 00:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "moderation"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "mistral-moderation-2603",
+    "name": "Mistral Moderation",
+    "provider": "mistral",
+    "family": "mistral-moderation",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "moderation"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "mistral-moderation-latest",
+    "name": "Mistral Moderation",
+    "provider": "mistral",
+    "family": "mistral-moderation",
+    "created_at": "2024-11-26 00:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "moderation"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
     }
   },
   {
@@ -8364,7 +23380,7 @@
     "name": "Mistral Nemo",
     "provider": "mistral",
     "family": "mistral-nemo",
-    "created_at": "2024-07-01 00:00:00 +0530",
+    "created_at": "2024-07-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": null,
@@ -8405,12 +23421,90 @@
       "knowledge": "2024-07"
     }
   },
+  {
+    "id": "mistral-ocr-2505",
+    "name": "Mistral Ocr 2505",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": "2025-05-22 23:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "vision"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "mistral-ocr-2512",
+    "name": "Mistral Ocr 2512",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "vision"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "mistral-ocr-latest",
+    "name": "Mistral Ocr Latest",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": "2025-05-22 23:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "vision"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
   {
     "id": "mistral-small-2506",
     "name": "Mistral Small 3.2",
     "provider": "mistral",
     "family": "mistral-small",
-    "created_at": "2025-06-20 00:00:00 +0530",
+    "created_at": "2025-06-20 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
@@ -8425,7 +23519,11 @@
     },
     "capabilities": [
       "function_calling",
-      "vision"
+      "vision",
+      "streaming",
+      "structured_output",
+      "batch",
+      "fine_tuning"
     ],
     "pricing": {
       "text_tokens": {
@@ -8436,6 +23534,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": true,
@@ -8458,7 +23558,7 @@
     "name": "Mistral Small 4",
     "provider": "mistral",
     "family": "mistral-small",
-    "created_at": "2026-03-16 00:00:00 +0530",
+    "created_at": "2026-03-16 00:00:00 UTC",
     "context_window": 256000,
     "max_output_tokens": 256000,
     "knowledge_cutoff": null,
@@ -8474,7 +23574,11 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming",
+      "structured_output",
+      "batch",
+      "fine_tuning"
     ],
     "pricing": {
       "text_tokens": {
@@ -8485,6 +23589,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": true,
@@ -8507,7 +23613,7 @@
     "name": "Mistral Small (latest)",
     "provider": "mistral",
     "family": "mistral-small",
-    "created_at": "2026-03-16 00:00:00 +0530",
+    "created_at": "2026-03-16 00:00:00 UTC",
     "context_window": 256000,
     "max_output_tokens": 256000,
     "knowledge_cutoff": null,
@@ -8523,7 +23629,11 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming",
+      "structured_output",
+      "batch",
+      "fine_tuning"
     ],
     "pricing": {
       "text_tokens": {
@@ -8534,6 +23644,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": true,
@@ -8551,12 +23663,157 @@
       "knowledge": "2025-06"
     }
   },
+  {
+    "id": "mistral-tiny-2407",
+    "name": "Mistral Tiny 2407",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": "2024-07-17 23:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "batch"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "mistral-tiny-latest",
+    "name": "Mistral Tiny Latest",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": "2024-07-17 23:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "batch"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "mistral-vibe-cli-fast",
+    "name": "Mistral Vibe Cli Fast",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "batch"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "mistral-vibe-cli-latest",
+    "name": "Mistral Vibe Cli Latest",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "batch"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "mistral-vibe-cli-with-tools",
+    "name": "Mistral Vibe Cli With Tools",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "batch"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
   {
     "id": "open-mistral-7b",
     "name": "Mistral 7B",
     "provider": "mistral",
     "family": "mistral",
-    "created_at": "2023-09-27 00:00:00 +0530",
+    "created_at": "2023-09-27 00:00:00 UTC",
     "context_window": 8000,
     "max_output_tokens": 8000,
     "knowledge_cutoff": null,
@@ -8597,12 +23854,70 @@
       "knowledge": "2023-12"
     }
   },
+  {
+    "id": "open-mistral-nemo",
+    "name": "Open Mistral Nemo",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": "2024-07-17 23:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "batch"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "open-mistral-nemo-2407",
+    "name": "Open Mistral Nemo 2407",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": "2024-07-17 23:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "batch"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
   {
     "id": "open-mixtral-8x22b",
     "name": "Mixtral 8x22B",
     "provider": "mistral",
     "family": "mixtral",
-    "created_at": "2024-04-17 00:00:00 +0530",
+    "created_at": "2024-04-17 00:00:00 UTC",
     "context_window": 64000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": null,
@@ -8648,7 +23963,7 @@
     "name": "Mixtral 8x7B",
     "provider": "mistral",
     "family": "mixtral",
-    "created_at": "2023-12-11 00:00:00 +0530",
+    "created_at": "2023-12-11 00:00:00 UTC",
     "context_window": 32000,
     "max_output_tokens": 32000,
     "knowledge_cutoff": null,
@@ -8694,7 +24009,7 @@
     "name": "Pixtral 12B",
     "provider": "mistral",
     "family": "pixtral",
-    "created_at": "2024-09-01 00:00:00 +0530",
+    "created_at": "2024-09-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": null,
@@ -8737,12 +24052,43 @@
       "knowledge": "2024-09"
     }
   },
+  {
+    "id": "pixtral-large-2411",
+    "name": "Pixtral Large",
+    "provider": "mistral",
+    "family": "pixtral",
+    "created_at": "2024-11-12 00:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "vision",
+      "batch"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
   {
     "id": "pixtral-large-latest",
     "name": "Pixtral Large (latest)",
     "provider": "mistral",
     "family": "pixtral",
-    "created_at": "2024-11-01 00:00:00 +0530",
+    "created_at": "2024-11-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": null,
@@ -8757,7 +24103,10 @@
     },
     "capabilities": [
       "function_calling",
-      "vision"
+      "vision",
+      "streaming",
+      "structured_output",
+      "batch"
     ],
     "pricing": {
       "text_tokens": {
@@ -8768,6 +24117,8 @@
       }
     },
     "metadata": {
+      "object": "model",
+      "owned_by": "mistralai",
       "source": "models.dev",
       "provider_id": "mistral",
       "open_weights": true,
@@ -8786,13 +24137,13 @@
     }
   },
   {
-    "id": "babbage-002",
-    "name": "Babbage 002",
-    "provider": "openai",
-    "family": "babbage",
-    "created_at": "2023-08-21 21:46:55 +0530",
-    "context_window": 4096,
-    "max_output_tokens": 16384,
+    "id": "voxtral-mini-2507",
+    "name": "Voxtral Mini 2507",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
     "knowledge_cutoff": null,
     "modalities": {
       "input": [
@@ -8805,6 +24156,286 @@
     "capabilities": [
       "streaming"
     ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "voxtral-mini-2602",
+    "name": "Voxtral Mini 2602",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "voxtral-mini-latest",
+    "name": "Voxtral Mini Latest",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "voxtral-mini-realtime-2602",
+    "name": "Voxtral Mini Realtime 2602",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "voxtral-mini-realtime-latest",
+    "name": "Voxtral Mini Realtime Latest",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "voxtral-mini-transcribe-2507",
+    "name": "Voxtral Mini Transcribe 2507",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "transcription"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "voxtral-mini-transcribe-realtime-2602",
+    "name": "Voxtral Mini Transcribe Realtime 2602",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "transcription"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "voxtral-mini-tts-2603",
+    "name": "Voxtral Mini Tts 2603",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "voxtral-mini-tts-latest",
+    "name": "Voxtral Mini Tts Latest",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "voxtral-small-2507",
+    "name": "Voxtral Small 2507",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "voxtral-small-latest",
+    "name": "Voxtral Small Latest",
+    "provider": "mistral",
+    "family": "mistral",
+    "created_at": null,
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "mistralai"
+    }
+  },
+  {
+    "id": "babbage-002",
+    "name": "babbage-002",
+    "provider": "openai",
+    "family": null,
+    "created_at": "2023-08-21 16:16:55 UTC",
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -8818,12 +24449,39 @@
       "owned_by": "system"
     }
   },
+  {
+    "id": "chat-latest",
+    "name": "chat-latest",
+    "provider": "openai",
+    "family": null,
+    "created_at": "2026-05-02 06:50:02 UTC",
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": "system"
+    }
+  },
   {
     "id": "chatgpt-image-latest",
     "name": "chatgpt-image-latest",
     "provider": "openai",
     "family": "gpt-image",
-    "created_at": "2025-12-16 00:00:00 +0530",
+    "created_at": "2025-12-16 00:00:00 UTC",
     "context_window": 0,
     "max_output_tokens": 0,
     "knowledge_cutoff": null,
@@ -8838,10 +24496,16 @@
       ]
     },
     "capabilities": [
-      "vision",
-      "streaming"
+      "vision"
     ],
-    "pricing": {},
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
     "metadata": {
       "object": "model",
       "owned_by": "system",
@@ -8858,75 +24522,20 @@
       }
     }
   },
-  {
-    "id": "codex-mini-latest",
-    "name": "Codex Mini",
-    "provider": "openai",
-    "family": "gpt-codex-mini",
-    "created_at": "2025-05-16 00:00:00 +0530",
-    "context_window": 200000,
-    "max_output_tokens": 100000,
-    "knowledge_cutoff": null,
-    "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
-    },
-    "capabilities": [
-      "function_calling",
-      "reasoning"
-    ],
-    "pricing": {
-      "text_tokens": {
-        "standard": {
-          "input_per_million": 1.5,
-          "output_per_million": 6,
-          "cached_input_per_million": 0.375
-        }
-      }
-    },
-    "metadata": {
-      "source": "models.dev",
-      "provider_id": "openai",
-      "open_weights": false,
-      "attachment": true,
-      "temperature": false,
-      "last_updated": "2025-05-16",
-      "cost": {
-        "input": 1.5,
-        "output": 6,
-        "cache_read": 0.375
-      },
-      "limit": {
-        "context": 200000,
-        "output": 100000
-      },
-      "knowledge": "2024-04"
-    }
-  },
   {
     "id": "computer-use-preview",
-    "name": "Computer Use Preview",
+    "name": "computer-use-preview",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2024-12-20 06:17:57 +0530",
+    "family": null,
+    "created_at": "2024-12-20 00:47:57 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -8942,24 +24551,18 @@
   },
   {
     "id": "computer-use-preview-2025-03-11",
-    "name": "Computer Use Preview 20250311",
+    "name": "computer-use-preview-2025-03-11",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-03-08 01:20:21 +0530",
+    "family": null,
+    "created_at": "2025-03-07 19:50:21 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -8975,25 +24578,18 @@
   },
   {
     "id": "dall-e-2",
-    "name": "DALL-E-2",
+    "name": "dall-e-2",
     "provider": "openai",
-    "family": "dall_e",
-    "created_at": "2023-11-01 05:52:57 +0530",
+    "family": null,
+    "created_at": "2023-11-01 00:22:57 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text",
-        "image"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -9009,25 +24605,18 @@
   },
   {
     "id": "dall-e-3",
-    "name": "DALL-E-3",
+    "name": "dall-e-3",
     "provider": "openai",
-    "family": "dall_e",
-    "created_at": "2023-11-01 02:16:29 +0530",
+    "family": null,
+    "created_at": "2023-10-31 20:46:29 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text",
-        "image"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -9043,24 +24632,18 @@
   },
   {
     "id": "davinci-002",
-    "name": "Davinci 002",
+    "name": "davinci-002",
     "provider": "openai",
-    "family": "davinci",
-    "created_at": "2023-08-21 21:41:41 +0530",
+    "family": null,
+    "created_at": "2023-08-21 16:11:41 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -9079,7 +24662,7 @@
     "name": "GPT-3.5-turbo",
     "provider": "openai",
     "family": "gpt",
-    "created_at": "2023-03-01 00:00:00 +0530",
+    "created_at": "2023-03-01 00:00:00 UTC",
     "context_window": 16385,
     "max_output_tokens": 4096,
     "knowledge_cutoff": "2021-09-01",
@@ -9091,15 +24674,13 @@
         "text"
       ]
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.5,
           "output_per_million": 1.5,
-          "cached_input_per_million": 1.25
+          "cache_read_input_per_million": 1.25
         }
       }
     },
@@ -9126,24 +24707,18 @@
   },
   {
     "id": "gpt-3.5-turbo-0125",
-    "name": "GPT-3.5 Turbo 0125",
+    "name": "gpt-3.5-turbo-0125",
     "provider": "openai",
-    "family": "gpt35_turbo",
-    "created_at": "2024-01-24 03:49:18 +0530",
+    "family": null,
+    "created_at": "2024-01-23 22:19:18 UTC",
     "context_window": 16385,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -9159,24 +24734,18 @@
   },
   {
     "id": "gpt-3.5-turbo-1106",
-    "name": "GPT-3.5 Turbo 1106",
+    "name": "gpt-3.5-turbo-1106",
     "provider": "openai",
-    "family": "gpt35_turbo",
-    "created_at": "2023-11-03 02:45:48 +0530",
+    "family": null,
+    "created_at": "2023-11-02 21:15:48 UTC",
     "context_window": 16385,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -9192,24 +24761,18 @@
   },
   {
     "id": "gpt-3.5-turbo-16k",
-    "name": "GPT-3.5 Turbo 16k",
+    "name": "gpt-3.5-turbo-16k",
     "provider": "openai",
-    "family": "gpt35_turbo",
-    "created_at": "2023-05-11 04:05:02 +0530",
+    "family": null,
+    "created_at": "2023-05-10 22:35:02 UTC",
     "context_window": 16385,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -9225,24 +24788,18 @@
   },
   {
     "id": "gpt-3.5-turbo-instruct",
-    "name": "GPT-3.5 Turbo Instruct",
+    "name": "gpt-3.5-turbo-instruct",
     "provider": "openai",
-    "family": "gpt35_turbo",
-    "created_at": "2023-08-24 23:53:47 +0530",
+    "family": null,
+    "created_at": "2023-08-24 18:23:47 UTC",
     "context_window": 16385,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -9258,24 +24815,18 @@
   },
   {
     "id": "gpt-3.5-turbo-instruct-0914",
-    "name": "GPT-3.5 Turbo Instruct 0914",
+    "name": "gpt-3.5-turbo-instruct-0914",
     "provider": "openai",
-    "family": "gpt35_turbo",
-    "created_at": "2023-09-08 03:04:32 +0530",
+    "family": null,
+    "created_at": "2023-09-07 21:34:32 UTC",
     "context_window": 16385,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -9294,7 +24845,7 @@
     "name": "GPT-4",
     "provider": "openai",
     "family": "gpt",
-    "created_at": "2023-11-06 00:00:00 +0530",
+    "created_at": "2023-11-06 00:00:00 UTC",
     "context_window": 8192,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -9308,7 +24859,7 @@
     },
     "capabilities": [
       "function_calling",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
@@ -9340,24 +24891,18 @@
   },
   {
     "id": "gpt-4-0613",
-    "name": "GPT-4 0613",
+    "name": "gpt-4-0613",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2023-06-12 22:24:56 +0530",
+    "family": null,
+    "created_at": "2023-06-12 16:54:56 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -9376,7 +24921,7 @@
     "name": "GPT-4 Turbo",
     "provider": "openai",
     "family": "gpt",
-    "created_at": "2023-11-06 00:00:00 +0530",
+    "created_at": "2023-11-06 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -9391,8 +24936,7 @@
     },
     "capabilities": [
       "function_calling",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
@@ -9424,26 +24968,20 @@
   },
   {
     "id": "gpt-4-turbo-2024-04-09",
-    "name": "GPT-4 Turbo 20240409",
+    "name": "gpt-4-turbo-2024-04-09",
     "provider": "openai",
-    "family": "gpt4_turbo",
-    "created_at": "2024-04-09 00:11:17 +0530",
+    "family": null,
+    "created_at": "2024-04-08 18:41:17 UTC",
     "context_window": 128000,
-    "max_output_tokens": 4096,
+    "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
-      "function_calling"
+      "function_calling",
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
@@ -9463,7 +25001,7 @@
     "name": "GPT-4.1",
     "provider": "openai",
     "family": "gpt",
-    "created_at": "2025-04-14 00:00:00 +0530",
+    "created_at": "2025-04-14 00:00:00 UTC",
     "context_window": 1047576,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -9480,15 +25018,14 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 2,
           "output_per_million": 8,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5
         }
       }
     },
@@ -9515,34 +25052,28 @@
   },
   {
     "id": "gpt-4.1-2025-04-14",
-    "name": "GPT-4.1 20250414",
+    "name": "gpt-4.1-2025-04-14",
     "provider": "openai",
-    "family": "gpt41",
-    "created_at": "2025-04-11 01:39:06 +0530",
+    "family": null,
+    "created_at": "2025-04-10 20:09:06 UTC",
     "context_window": 1047576,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 2.0,
           "output_per_million": 8.0,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5
         }
       }
     },
@@ -9556,7 +25087,7 @@
     "name": "GPT-4.1 mini",
     "provider": "openai",
     "family": "gpt-mini",
-    "created_at": "2025-04-14 00:00:00 +0530",
+    "created_at": "2025-04-14 00:00:00 UTC",
     "context_window": 1047576,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -9573,15 +25104,14 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.4,
           "output_per_million": 1.6,
-          "cached_input_per_million": 0.1
+          "cache_read_input_per_million": 0.1
         }
       }
     },
@@ -9608,34 +25138,28 @@
   },
   {
     "id": "gpt-4.1-mini-2025-04-14",
-    "name": "GPT-4.1 Mini 20250414",
+    "name": "gpt-4.1-mini-2025-04-14",
     "provider": "openai",
-    "family": "gpt41_mini",
-    "created_at": "2025-04-11 02:09:07 +0530",
+    "family": null,
+    "created_at": "2025-04-10 20:39:07 UTC",
     "context_window": 1047576,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.4,
           "output_per_million": 1.6,
-          "cached_input_per_million": 0.1
+          "cache_read_input_per_million": 0.1
         }
       }
     },
@@ -9649,7 +25173,7 @@
     "name": "GPT-4.1 nano",
     "provider": "openai",
     "family": "gpt-nano",
-    "created_at": "2025-04-14 00:00:00 +0530",
+    "created_at": "2025-04-14 00:00:00 UTC",
     "context_window": 1047576,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -9665,15 +25189,14 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.1,
           "output_per_million": 0.4,
-          "cached_input_per_million": 0.03
+          "cache_read_input_per_million": 0.03
         }
       }
     },
@@ -9700,27 +25223,21 @@
   },
   {
     "id": "gpt-4.1-nano-2025-04-14",
-    "name": "GPT-4.1 Nano 20250414",
+    "name": "gpt-4.1-nano-2025-04-14",
     "provider": "openai",
-    "family": "gpt41_nano",
-    "created_at": "2025-04-11 03:07:05 +0530",
+    "family": null,
+    "created_at": "2025-04-10 21:37:05 UTC",
     "context_window": 1047576,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
@@ -9740,7 +25257,7 @@
     "name": "GPT-4o",
     "provider": "openai",
     "family": "gpt",
-    "created_at": "2024-05-13 00:00:00 +0530",
+    "created_at": "2024-05-13 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
@@ -9757,15 +25274,14 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 2.5,
           "output_per_million": 10,
-          "cached_input_per_million": 1.25
+          "cache_read_input_per_million": 1.25
         }
       }
     },
@@ -9795,7 +25311,7 @@
     "name": "GPT-4o (2024-05-13)",
     "provider": "openai",
     "family": "gpt",
-    "created_at": "2024-05-13 00:00:00 +0530",
+    "created_at": "2024-05-13 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -9811,8 +25327,7 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
@@ -9847,7 +25362,7 @@
     "name": "GPT-4o (2024-08-06)",
     "provider": "openai",
     "family": "gpt",
-    "created_at": "2024-08-06 00:00:00 +0530",
+    "created_at": "2024-08-06 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
@@ -9863,15 +25378,14 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 2.5,
           "output_per_million": 10,
-          "cached_input_per_million": 1.25
+          "cache_read_input_per_million": 1.25
         }
       }
     },
@@ -9901,7 +25415,7 @@
     "name": "GPT-4o (2024-11-20)",
     "provider": "openai",
     "family": "gpt",
-    "created_at": "2024-11-20 00:00:00 +0530",
+    "created_at": "2024-11-20 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
@@ -9917,15 +25431,14 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 2.5,
           "output_per_million": 10,
-          "cached_input_per_million": 1.25
+          "cache_read_input_per_million": 1.25
         }
       }
     },
@@ -9952,28 +25465,18 @@
   },
   {
     "id": "gpt-4o-audio-preview",
-    "name": "GPT-4o-Audio Preview",
+    "name": "gpt-4o-audio-preview",
     "provider": "openai",
-    "family": "gpt4o_audio",
-    "created_at": "2024-09-27 23:37:23 +0530",
+    "family": null,
+    "created_at": "2024-09-27 18:07:23 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming",
-      "speech_generation",
-      "transcription"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -9989,28 +25492,18 @@
   },
   {
     "id": "gpt-4o-audio-preview-2024-12-17",
-    "name": "GPT-4o-Audio Preview 20241217",
+    "name": "gpt-4o-audio-preview-2024-12-17",
     "provider": "openai",
-    "family": "gpt4o_audio",
-    "created_at": "2024-12-13 01:40:39 +0530",
+    "family": null,
+    "created_at": "2024-12-12 20:10:39 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming",
-      "speech_generation",
-      "transcription"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10026,28 +25519,18 @@
   },
   {
     "id": "gpt-4o-audio-preview-2025-06-03",
-    "name": "GPT-4o-Audio Preview 20250603",
+    "name": "gpt-4o-audio-preview-2025-06-03",
     "provider": "openai",
-    "family": "gpt4o_audio",
-    "created_at": "2025-06-03 05:24:58 +0530",
+    "family": null,
+    "created_at": "2025-06-02 23:54:58 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming",
-      "speech_generation",
-      "transcription"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10066,7 +25549,7 @@
     "name": "GPT-4o mini",
     "provider": "openai",
     "family": "gpt-mini",
-    "created_at": "2024-07-18 00:00:00 +0530",
+    "created_at": "2024-07-18 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
@@ -10083,15 +25566,14 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.15,
           "output_per_million": 0.6,
-          "cached_input_per_million": 0.08
+          "cache_read_input_per_million": 0.08
         }
       }
     },
@@ -10118,27 +25600,21 @@
   },
   {
     "id": "gpt-4o-mini-2024-07-18",
-    "name": "GPT-4o-Mini 20240718",
+    "name": "gpt-4o-mini-2024-07-18",
     "provider": "openai",
-    "family": "gpt4o_mini",
-    "created_at": "2024-07-17 05:01:57 +0530",
+    "family": null,
+    "created_at": "2024-07-16 23:31:57 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
@@ -10155,28 +25631,18 @@
   },
   {
     "id": "gpt-4o-mini-audio-preview",
-    "name": "GPT-4o-Mini Audio Preview",
+    "name": "gpt-4o-mini-audio-preview",
     "provider": "openai",
-    "family": "gpt4o_mini_audio",
-    "created_at": "2024-12-17 03:47:04 +0530",
+    "family": null,
+    "created_at": "2024-12-16 22:17:04 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming",
-      "speech_generation",
-      "transcription"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10192,28 +25658,18 @@
   },
   {
     "id": "gpt-4o-mini-audio-preview-2024-12-17",
-    "name": "GPT-4o-Mini Audio Preview 20241217",
+    "name": "gpt-4o-mini-audio-preview-2024-12-17",
     "provider": "openai",
-    "family": "gpt4o_mini_audio",
-    "created_at": "2024-12-14 00:22:00 +0530",
+    "family": null,
+    "created_at": "2024-12-13 18:52:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming",
-      "speech_generation",
-      "transcription"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10229,24 +25685,18 @@
   },
   {
     "id": "gpt-4o-mini-realtime-preview",
-    "name": "GPT-4o-Mini Realtime Preview",
+    "name": "gpt-4o-mini-realtime-preview",
     "provider": "openai",
-    "family": "gpt4o_mini_realtime",
-    "created_at": "2024-12-17 03:46:20 +0530",
+    "family": null,
+    "created_at": "2024-12-16 22:16:20 UTC",
     "context_window": 128000,
-    "max_output_tokens": 4096,
+    "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10262,24 +25712,18 @@
   },
   {
     "id": "gpt-4o-mini-realtime-preview-2024-12-17",
-    "name": "GPT-4o-Mini Realtime Preview 20241217",
+    "name": "gpt-4o-mini-realtime-preview-2024-12-17",
     "provider": "openai",
-    "family": "gpt4o_mini_realtime",
-    "created_at": "2024-12-13 23:26:41 +0530",
+    "family": null,
+    "created_at": "2024-12-13 17:56:41 UTC",
     "context_window": 128000,
-    "max_output_tokens": 4096,
+    "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10295,24 +25739,18 @@
   },
   {
     "id": "gpt-4o-mini-search-preview",
-    "name": "GPT-4o-Mini Search Preview",
+    "name": "gpt-4o-mini-search-preview",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-03-08 05:16:01 +0530",
+    "family": null,
+    "created_at": "2025-03-07 23:46:01 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10328,24 +25766,18 @@
   },
   {
     "id": "gpt-4o-mini-search-preview-2025-03-11",
-    "name": "GPT-4o-Mini Search Preview 20250311",
+    "name": "gpt-4o-mini-search-preview-2025-03-11",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-03-08 05:10:58 +0530",
+    "family": null,
+    "created_at": "2025-03-07 23:40:58 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10361,25 +25793,18 @@
   },
   {
     "id": "gpt-4o-mini-transcribe",
-    "name": "GPT-4o-Mini Transcribe",
+    "name": "gpt-4o-mini-transcribe",
     "provider": "openai",
-    "family": "gpt4o_mini_transcribe",
-    "created_at": "2025-03-16 01:26:36 +0530",
+    "family": null,
+    "created_at": "2025-03-15 19:56:36 UTC",
     "context_window": 16000,
     "max_output_tokens": 2000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10395,25 +25820,18 @@
   },
   {
     "id": "gpt-4o-mini-transcribe-2025-03-20",
-    "name": "GPT-4o-Mini Transcribe 20250320",
+    "name": "gpt-4o-mini-transcribe-2025-03-20",
     "provider": "openai",
-    "family": "gpt4o_mini_transcribe",
-    "created_at": "2025-12-13 12:52:25 +0530",
+    "family": null,
+    "created_at": "2025-12-13 07:22:25 UTC",
     "context_window": 16000,
     "max_output_tokens": 2000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10429,25 +25847,18 @@
   },
   {
     "id": "gpt-4o-mini-transcribe-2025-12-15",
-    "name": "GPT-4o-Mini Transcribe 20251215",
+    "name": "gpt-4o-mini-transcribe-2025-12-15",
     "provider": "openai",
-    "family": "gpt4o_mini_transcribe",
-    "created_at": "2025-12-13 12:50:07 +0530",
+    "family": null,
+    "created_at": "2025-12-13 07:20:07 UTC",
     "context_window": 16000,
     "max_output_tokens": 2000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10463,26 +25874,18 @@
   },
   {
     "id": "gpt-4o-mini-tts",
-    "name": "GPT-4o-Mini Tts",
+    "name": "gpt-4o-mini-tts",
     "provider": "openai",
-    "family": "gpt4o_mini_tts",
-    "created_at": "2025-03-19 22:35:59 +0530",
+    "family": null,
+    "created_at": "2025-03-19 17:05:59 UTC",
     "context_window": null,
     "max_output_tokens": null,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10498,26 +25901,18 @@
   },
   {
     "id": "gpt-4o-mini-tts-2025-03-20",
-    "name": "GPT-4o-Mini Tts 20250320",
+    "name": "gpt-4o-mini-tts-2025-03-20",
     "provider": "openai",
-    "family": "gpt4o_mini_tts",
-    "created_at": "2025-12-13 12:55:31 +0530",
+    "family": null,
+    "created_at": "2025-12-13 07:25:31 UTC",
     "context_window": null,
     "max_output_tokens": null,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10533,26 +25928,18 @@
   },
   {
     "id": "gpt-4o-mini-tts-2025-12-15",
-    "name": "GPT-4o-Mini Tts 20251215",
+    "name": "gpt-4o-mini-tts-2025-12-15",
     "provider": "openai",
-    "family": "gpt4o_mini_tts",
-    "created_at": "2025-12-13 12:57:17 +0530",
+    "family": null,
+    "created_at": "2025-12-13 07:27:17 UTC",
     "context_window": null,
     "max_output_tokens": null,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10568,24 +25955,18 @@
   },
   {
     "id": "gpt-4o-realtime-preview",
-    "name": "GPT-4o-Realtime Preview",
+    "name": "gpt-4o-realtime-preview",
     "provider": "openai",
-    "family": "gpt4o_realtime",
-    "created_at": "2024-09-30 07:03:18 +0530",
+    "family": null,
+    "created_at": "2024-09-30 01:33:18 UTC",
     "context_window": 128000,
-    "max_output_tokens": 4096,
+    "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10601,24 +25982,18 @@
   },
   {
     "id": "gpt-4o-realtime-preview-2024-12-17",
-    "name": "GPT-4o-Realtime Preview 20241217",
+    "name": "gpt-4o-realtime-preview-2024-12-17",
     "provider": "openai",
-    "family": "gpt4o_realtime",
-    "created_at": "2024-12-12 01:00:30 +0530",
+    "family": null,
+    "created_at": "2024-12-11 19:30:30 UTC",
     "context_window": 128000,
-    "max_output_tokens": 4096,
+    "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10634,24 +26009,18 @@
   },
   {
     "id": "gpt-4o-realtime-preview-2025-06-03",
-    "name": "GPT-4o-Realtime Preview 20250603",
+    "name": "gpt-4o-realtime-preview-2025-06-03",
     "provider": "openai",
-    "family": "gpt4o_realtime",
-    "created_at": "2025-06-03 05:13:58 +0530",
+    "family": null,
+    "created_at": "2025-06-02 23:43:58 UTC",
     "context_window": 128000,
-    "max_output_tokens": 4096,
+    "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10667,25 +26036,19 @@
   },
   {
     "id": "gpt-4o-search-preview",
-    "name": "GPT-4o Search Preview",
+    "name": "gpt-4o-search-preview",
     "provider": "openai",
-    "family": "gpt4o_search",
-    "created_at": "2026-02-24 09:28:54 +0530",
+    "family": null,
+    "created_at": "2026-02-24 03:58:54 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
@@ -10702,25 +26065,19 @@
   },
   {
     "id": "gpt-4o-search-preview-2025-03-11",
-    "name": "GPT-4o Search Preview 20250311",
+    "name": "gpt-4o-search-preview-2025-03-11",
     "provider": "openai",
-    "family": "gpt4o_search",
-    "created_at": "2026-02-24 09:30:21 +0530",
+    "family": null,
+    "created_at": "2026-02-24 04:00:21 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
@@ -10737,25 +26094,18 @@
   },
   {
     "id": "gpt-4o-transcribe",
-    "name": "GPT-4o-Transcribe",
+    "name": "gpt-4o-transcribe",
     "provider": "openai",
-    "family": "gpt4o_transcribe",
-    "created_at": "2025-03-16 01:24:23 +0530",
+    "family": null,
+    "created_at": "2025-03-15 19:54:23 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10771,25 +26121,18 @@
   },
   {
     "id": "gpt-4o-transcribe-diarize",
-    "name": "GPT-4o-Transcribe Diarize",
+    "name": "gpt-4o-transcribe-diarize",
     "provider": "openai",
-    "family": "gpt4o_transcribe",
-    "created_at": "2025-06-25 02:31:27 +0530",
+    "family": null,
+    "created_at": "2025-06-24 21:01:27 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -10808,7 +26151,7 @@
     "name": "GPT-5",
     "provider": "openai",
     "family": "gpt",
-    "created_at": "2025-08-07 00:00:00 +0530",
+    "created_at": "2025-08-07 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-09-30",
@@ -10825,15 +26168,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
         }
       }
     },
@@ -10861,27 +26203,21 @@
   },
   {
     "id": "gpt-5-2025-08-07",
-    "name": "GPT-5 20250807",
+    "name": "gpt-5-2025-08-07",
     "provider": "openai",
-    "family": "gpt5",
-    "created_at": "2025-08-02 00:39:20 +0530",
+    "family": null,
+    "created_at": "2025-08-01 19:09:20 UTC",
     "context_window": 128000,
     "max_output_tokens": 400000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
       "structured_output",
+      "vision",
       "reasoning"
     ],
     "pricing": {
@@ -10889,7 +26225,7 @@
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10.0,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
         }
       }
     },
@@ -10903,7 +26239,7 @@
     "name": "GPT-5 Chat (latest)",
     "provider": "openai",
     "family": "gpt-codex",
-    "created_at": "2025-08-07 00:00:00 +0530",
+    "created_at": "2025-08-07 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-09-30",
@@ -10920,7 +26256,6 @@
       "structured_output",
       "reasoning",
       "vision",
-      "streaming",
       "function_calling"
     ],
     "pricing": {
@@ -10957,7 +26292,7 @@
     "name": "GPT-5-Codex",
     "provider": "openai",
     "family": "gpt-codex",
-    "created_at": "2025-09-15 00:00:00 +0530",
+    "created_at": "2025-09-15 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-09-30",
@@ -10974,15 +26309,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
         }
       }
     },
@@ -11013,7 +26347,7 @@
     "name": "GPT-5 Mini",
     "provider": "openai",
     "family": "gpt-mini",
-    "created_at": "2025-08-07 00:00:00 +0530",
+    "created_at": "2025-08-07 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-05-30",
@@ -11030,15 +26364,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.25,
           "output_per_million": 2,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       }
     },
@@ -11066,35 +26399,29 @@
   },
   {
     "id": "gpt-5-mini-2025-08-07",
-    "name": "GPT-5 Mini 20250807",
+    "name": "gpt-5-mini-2025-08-07",
     "provider": "openai",
-    "family": "gpt5",
-    "created_at": "2025-08-06 02:01:07 +0530",
+    "family": null,
+    "created_at": "2025-08-05 20:31:07 UTC",
     "context_window": 128000,
     "max_output_tokens": 400000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
       "structured_output",
+      "vision",
       "reasoning"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
-          "input_per_million": 1.25,
-          "output_per_million": 10.0,
-          "cached_input_per_million": 0.125
+          "input_per_million": 0.25,
+          "output_per_million": 2.0,
+          "cache_read_input_per_million": 0.025
         }
       }
     },
@@ -11108,7 +26435,7 @@
     "name": "GPT-5 Nano",
     "provider": "openai",
     "family": "gpt-nano",
-    "created_at": "2025-08-07 00:00:00 +0530",
+    "created_at": "2025-08-07 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-05-30",
@@ -11125,15 +26452,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.05,
           "output_per_million": 0.4,
-          "cached_input_per_million": 0.005
+          "cache_read_input_per_million": 0.005
         }
       }
     },
@@ -11161,35 +26487,29 @@
   },
   {
     "id": "gpt-5-nano-2025-08-07",
-    "name": "GPT-5 Nano 20250807",
+    "name": "gpt-5-nano-2025-08-07",
     "provider": "openai",
-    "family": "gpt5",
-    "created_at": "2025-08-06 02:08:23 +0530",
+    "family": null,
+    "created_at": "2025-08-05 20:38:23 UTC",
     "context_window": 128000,
     "max_output_tokens": 400000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
       "structured_output",
+      "vision",
       "reasoning"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
-          "input_per_million": 1.25,
-          "output_per_million": 10.0,
-          "cached_input_per_million": 0.125
+          "input_per_million": 0.05,
+          "output_per_million": 0.4,
+          "cache_read_input_per_million": 0.005
         }
       }
     },
@@ -11203,7 +26523,7 @@
     "name": "GPT-5 Pro",
     "provider": "openai",
     "family": "gpt-pro",
-    "created_at": "2025-10-06 00:00:00 +0530",
+    "created_at": "2025-10-06 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 272000,
     "knowledge_cutoff": "2024-09-30",
@@ -11220,8 +26540,7 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
@@ -11254,27 +26573,21 @@
   },
   {
     "id": "gpt-5-pro-2025-10-06",
-    "name": "GPT-5 Pro 20251006",
+    "name": "gpt-5-pro-2025-10-06",
     "provider": "openai",
-    "family": "gpt5",
-    "created_at": "2025-10-03 11:05:07 +0530",
+    "family": null,
+    "created_at": "2025-10-03 05:35:07 UTC",
     "context_window": 128000,
     "max_output_tokens": 400000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
       "structured_output",
+      "vision",
       "reasoning"
     ],
     "pricing": {
@@ -11282,7 +26595,7 @@
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10.0,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
         }
       }
     },
@@ -11293,27 +26606,21 @@
   },
   {
     "id": "gpt-5-search-api",
-    "name": "GPT-5 Search Api",
+    "name": "gpt-5-search-api",
     "provider": "openai",
-    "family": "gpt5",
-    "created_at": "2025-10-03 23:33:49 +0530",
+    "family": null,
+    "created_at": "2025-10-03 18:03:49 UTC",
     "context_window": 128000,
     "max_output_tokens": 400000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
       "structured_output",
+      "vision",
       "reasoning"
     ],
     "pricing": {
@@ -11321,7 +26628,7 @@
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10.0,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
         }
       }
     },
@@ -11332,27 +26639,21 @@
   },
   {
     "id": "gpt-5-search-api-2025-10-14",
-    "name": "GPT-5 Search Api 20251014",
+    "name": "gpt-5-search-api-2025-10-14",
     "provider": "openai",
-    "family": "gpt5",
-    "created_at": "2025-10-10 02:36:00 +0530",
+    "family": null,
+    "created_at": "2025-10-09 21:06:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 400000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
       "structured_output",
+      "vision",
       "reasoning"
     ],
     "pricing": {
@@ -11360,7 +26661,7 @@
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10.0,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
         }
       }
     },
@@ -11374,7 +26675,7 @@
     "name": "GPT-5.1",
     "provider": "openai",
     "family": "gpt",
-    "created_at": "2025-11-13 00:00:00 +0530",
+    "created_at": "2025-11-13 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-09-30",
@@ -11391,15 +26692,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.13
+          "cache_read_input_per_million": 0.13
         }
       }
     },
@@ -11427,27 +26727,21 @@
   },
   {
     "id": "gpt-5.1-2025-11-13",
-    "name": "GPT-5.1 20251113",
+    "name": "gpt-5.1-2025-11-13",
     "provider": "openai",
-    "family": "gpt5",
-    "created_at": "2025-11-11 00:15:53 +0530",
+    "family": null,
+    "created_at": "2025-11-10 18:45:53 UTC",
     "context_window": 128000,
     "max_output_tokens": 400000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
       "structured_output",
+      "vision",
       "reasoning"
     ],
     "pricing": {
@@ -11455,7 +26749,7 @@
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10.0,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
         }
       }
     },
@@ -11469,7 +26763,7 @@
     "name": "GPT-5.1 Chat",
     "provider": "openai",
     "family": "gpt-codex",
-    "created_at": "2025-11-13 00:00:00 +0530",
+    "created_at": "2025-11-13 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": "2024-09-30",
@@ -11486,15 +26780,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
         }
       }
     },
@@ -11524,7 +26817,7 @@
     "name": "GPT-5.1 Codex",
     "provider": "openai",
     "family": "gpt-codex",
-    "created_at": "2025-11-13 00:00:00 +0530",
+    "created_at": "2025-11-13 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-09-30",
@@ -11541,15 +26834,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
         }
       }
     },
@@ -11580,7 +26872,7 @@
     "name": "GPT-5.1 Codex Max",
     "provider": "openai",
     "family": "gpt-codex",
-    "created_at": "2025-11-13 00:00:00 +0530",
+    "created_at": "2025-11-13 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-09-30",
@@ -11597,15 +26889,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
         }
       }
     },
@@ -11636,7 +26927,7 @@
     "name": "GPT-5.1 Codex mini",
     "provider": "openai",
     "family": "gpt-codex",
-    "created_at": "2025-11-13 00:00:00 +0530",
+    "created_at": "2025-11-13 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-09-30",
@@ -11653,15 +26944,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.25,
           "output_per_million": 2,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       }
     },
@@ -11692,7 +26982,7 @@
     "name": "GPT-5.2",
     "provider": "openai",
     "family": "gpt",
-    "created_at": "2025-12-11 00:00:00 +0530",
+    "created_at": "2025-12-11 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2025-08-31",
@@ -11709,15 +26999,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.75,
           "output_per_million": 14,
-          "cached_input_per_million": 0.175
+          "cache_read_input_per_million": 0.175
         }
       }
     },
@@ -11745,27 +27034,21 @@
   },
   {
     "id": "gpt-5.2-2025-12-11",
-    "name": "GPT-5.2 20251211",
+    "name": "gpt-5.2-2025-12-11",
     "provider": "openai",
-    "family": "gpt5",
-    "created_at": "2025-12-10 02:13:48 +0530",
+    "family": null,
+    "created_at": "2025-12-09 20:43:48 UTC",
     "context_window": 128000,
     "max_output_tokens": 400000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
       "structured_output",
+      "vision",
       "reasoning"
     ],
     "pricing": {
@@ -11773,7 +27056,7 @@
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10.0,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
         }
       }
     },
@@ -11787,7 +27070,7 @@
     "name": "GPT-5.2 Chat",
     "provider": "openai",
     "family": "gpt-codex",
-    "created_at": "2025-12-11 00:00:00 +0530",
+    "created_at": "2025-12-11 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": "2025-08-31",
@@ -11804,15 +27087,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.75,
           "output_per_million": 14,
-          "cached_input_per_million": 0.175
+          "cache_read_input_per_million": 0.175
         }
       }
     },
@@ -11842,7 +27124,7 @@
     "name": "GPT-5.2 Codex",
     "provider": "openai",
     "family": "gpt-codex",
-    "created_at": "2025-12-11 00:00:00 +0530",
+    "created_at": "2025-12-11 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2025-08-31",
@@ -11860,15 +27142,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.75,
           "output_per_million": 14,
-          "cached_input_per_million": 0.175
+          "cache_read_input_per_million": 0.175
         }
       }
     },
@@ -11899,7 +27180,7 @@
     "name": "GPT-5.2 Pro",
     "provider": "openai",
     "family": "gpt-pro",
-    "created_at": "2025-12-11 00:00:00 +0530",
+    "created_at": "2025-12-11 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2025-08-31",
@@ -11916,7 +27197,6 @@
       "function_calling",
       "reasoning",
       "vision",
-      "streaming",
       "structured_output"
     ],
     "pricing": {
@@ -11950,27 +27230,21 @@
   },
   {
     "id": "gpt-5.2-pro-2025-12-11",
-    "name": "GPT-5.2 Pro 20251211",
+    "name": "gpt-5.2-pro-2025-12-11",
     "provider": "openai",
-    "family": "gpt5",
-    "created_at": "2025-12-10 10:49:19 +0530",
+    "family": null,
+    "created_at": "2025-12-10 05:19:19 UTC",
     "context_window": 128000,
     "max_output_tokens": 400000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
       "structured_output",
+      "vision",
       "reasoning"
     ],
     "pricing": {
@@ -11978,7 +27252,7 @@
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10.0,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
         }
       }
     },
@@ -11992,7 +27266,7 @@
     "name": "GPT-5.3 Chat (latest)",
     "provider": "openai",
     "family": "gpt",
-    "created_at": "2026-03-03 00:00:00 +0530",
+    "created_at": "2026-03-03 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": "2025-08-31",
@@ -12009,7 +27283,6 @@
       "function_calling",
       "structured_output",
       "vision",
-      "streaming",
       "reasoning"
     ],
     "pricing": {
@@ -12017,7 +27290,7 @@
         "standard": {
           "input_per_million": 1.75,
           "output_per_million": 14,
-          "cached_input_per_million": 0.175
+          "cache_read_input_per_million": 0.175
         }
       }
     },
@@ -12047,7 +27320,7 @@
     "name": "GPT-5.3 Codex",
     "provider": "openai",
     "family": "gpt-codex",
-    "created_at": "2026-02-05 00:00:00 +0530",
+    "created_at": "2026-02-05 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2025-08-31",
@@ -12065,15 +27338,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.75,
           "output_per_million": 14,
-          "cached_input_per_million": 0.175
+          "cache_read_input_per_million": 0.175
         }
       }
     },
@@ -12104,7 +27376,7 @@
     "name": "GPT-5.3 Codex Spark",
     "provider": "openai",
     "family": "gpt-codex-spark",
-    "created_at": "2026-02-05 00:00:00 +0530",
+    "created_at": "2026-02-05 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 32000,
     "knowledge_cutoff": "2025-08-31",
@@ -12129,7 +27401,7 @@
         "standard": {
           "input_per_million": 1.75,
           "output_per_million": 14,
-          "cached_input_per_million": 0.175
+          "cache_read_input_per_million": 0.175
         }
       }
     },
@@ -12158,7 +27430,7 @@
     "name": "GPT-5.4",
     "provider": "openai",
     "family": "gpt",
-    "created_at": "2026-03-05 00:00:00 +0530",
+    "created_at": "2026-03-05 00:00:00 UTC",
     "context_window": 1050000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2025-08-31",
@@ -12176,15 +27448,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 2.5,
           "output_per_million": 15,
-          "cached_input_per_million": 0.25
+          "cache_read_input_per_million": 0.25
         }
       }
     },
@@ -12217,27 +27488,21 @@
   },
   {
     "id": "gpt-5.4-2026-03-05",
-    "name": "GPT-5.4 20260305",
+    "name": "gpt-5.4-2026-03-05",
     "provider": "openai",
-    "family": "gpt5",
-    "created_at": "2026-03-05 01:24:22 +0530",
+    "family": null,
+    "created_at": "2026-03-04 19:54:22 UTC",
     "context_window": 128000,
     "max_output_tokens": 400000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
       "structured_output",
+      "vision",
       "reasoning"
     ],
     "pricing": {
@@ -12245,7 +27510,7 @@
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10.0,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
         }
       }
     },
@@ -12259,7 +27524,7 @@
     "name": "GPT-5.4 mini",
     "provider": "openai",
     "family": "gpt-mini",
-    "created_at": "2026-03-17 00:00:00 +0530",
+    "created_at": "2026-03-17 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2025-08-31",
@@ -12276,15 +27541,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.75,
           "output_per_million": 4.5,
-          "cached_input_per_million": 0.075
+          "cache_read_input_per_million": 0.075
         }
       }
     },
@@ -12312,35 +27576,29 @@
   },
   {
     "id": "gpt-5.4-mini-2026-03-17",
-    "name": "GPT-5.4 Mini 20260317",
+    "name": "gpt-5.4-mini-2026-03-17",
     "provider": "openai",
-    "family": "gpt5",
-    "created_at": "2026-03-14 06:47:56 +0530",
+    "family": null,
+    "created_at": "2026-03-14 01:17:56 UTC",
     "context_window": 128000,
     "max_output_tokens": 400000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
       "structured_output",
+      "vision",
       "reasoning"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
-          "input_per_million": 1.25,
-          "output_per_million": 10.0,
-          "cached_input_per_million": 0.125
+          "input_per_million": 0.25,
+          "output_per_million": 2.0,
+          "cache_read_input_per_million": 0.025
         }
       }
     },
@@ -12354,7 +27612,7 @@
     "name": "GPT-5.4 nano",
     "provider": "openai",
     "family": "gpt-nano",
-    "created_at": "2026-03-17 00:00:00 +0530",
+    "created_at": "2026-03-17 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2025-08-31",
@@ -12371,15 +27629,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.2,
           "output_per_million": 1.25,
-          "cached_input_per_million": 0.02
+          "cache_read_input_per_million": 0.02
         }
       }
     },
@@ -12407,35 +27664,29 @@
   },
   {
     "id": "gpt-5.4-nano-2026-03-17",
-    "name": "GPT-5.4 Nano 20260317",
+    "name": "gpt-5.4-nano-2026-03-17",
     "provider": "openai",
-    "family": "gpt5",
-    "created_at": "2026-03-14 06:43:57 +0530",
+    "family": null,
+    "created_at": "2026-03-14 01:13:57 UTC",
     "context_window": 128000,
     "max_output_tokens": 400000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
       "structured_output",
+      "vision",
       "reasoning"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
-          "input_per_million": 1.25,
-          "output_per_million": 10.0,
-          "cached_input_per_million": 0.125
+          "input_per_million": 0.05,
+          "output_per_million": 0.4,
+          "cache_read_input_per_million": 0.005
         }
       }
     },
@@ -12449,7 +27700,7 @@
     "name": "GPT-5.4 Pro",
     "provider": "openai",
     "family": "gpt-pro",
-    "created_at": "2026-03-05 00:00:00 +0530",
+    "created_at": "2026-03-05 00:00:00 UTC",
     "context_window": 1050000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2025-08-31",
@@ -12466,7 +27717,6 @@
       "function_calling",
       "reasoning",
       "vision",
-      "streaming",
       "structured_output"
     ],
     "pricing": {
@@ -12504,13 +27754,46 @@
   },
   {
     "id": "gpt-5.4-pro-2026-03-05",
-    "name": "GPT-5.4 Pro 20260305",
+    "name": "gpt-5.4-pro-2026-03-05",
     "provider": "openai",
-    "family": "gpt5",
-    "created_at": "2026-03-05 02:57:37 +0530",
+    "family": null,
+    "created_at": "2026-03-04 21:27:37 UTC",
     "context_window": 128000,
     "max_output_tokens": 400000,
     "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": "system"
+    }
+  },
+  {
+    "id": "gpt-5.5",
+    "name": "GPT-5.5",
+    "provider": "openai",
+    "family": "gpt",
+    "created_at": "2026-04-23 00:00:00 UTC",
+    "context_window": 1050000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": "2025-12-01",
     "modalities": {
       "input": [
         "text",
@@ -12522,9 +27805,64 @@
       ]
     },
     "capabilities": [
-      "streaming",
       "function_calling",
       "structured_output",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5,
+          "output_per_million": 30,
+          "cache_read_input_per_million": 0.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": "system",
+      "source": "models.dev",
+      "provider_id": "openai",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": false,
+      "last_updated": "2026-04-23",
+      "cost": {
+        "input": 5,
+        "output": 30,
+        "cache_read": 0.5,
+        "context_over_200k": {
+          "input": 10,
+          "output": 45,
+          "cache_read": 1
+        }
+      },
+      "limit": {
+        "context": 1050000,
+        "input": 922000,
+        "output": 128000
+      },
+      "knowledge": "2025-12-01"
+    }
+  },
+  {
+    "id": "gpt-5.5-2026-04-23",
+    "name": "gpt-5.5-2026-04-23",
+    "provider": "openai",
+    "family": null,
+    "created_at": "2026-04-22 06:27:21 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
       "reasoning"
     ],
     "pricing": {
@@ -12532,7 +27870,98 @@
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10.0,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": "system"
+    }
+  },
+  {
+    "id": "gpt-5.5-pro",
+    "name": "GPT-5.5 Pro",
+    "provider": "openai",
+    "family": "gpt-pro",
+    "created_at": "2026-04-23 00:00:00 UTC",
+    "context_window": 1050000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": "2025-12-01",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 30,
+          "output_per_million": 180
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": "system",
+      "source": "models.dev",
+      "provider_id": "openai",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": false,
+      "last_updated": "2026-04-23",
+      "cost": {
+        "input": 30,
+        "output": 180,
+        "context_over_200k": {
+          "input": 60,
+          "output": 270
+        }
+      },
+      "limit": {
+        "context": 1050000,
+        "input": 922000,
+        "output": 128000
+      },
+      "knowledge": "2025-12-01"
+    }
+  },
+  {
+    "id": "gpt-5.5-pro-2026-04-23",
+    "name": "gpt-5.5-pro-2026-04-23",
+    "provider": "openai",
+    "family": null,
+    "created_at": "2026-04-22 21:47:50 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 400000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125
         }
       }
     },
@@ -12543,26 +27972,18 @@
   },
   {
     "id": "gpt-audio",
-    "name": "GPT-Audio",
+    "name": "gpt-audio",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-08-28 05:30:49 +0530",
+    "family": null,
+    "created_at": "2025-08-28 00:00:49 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -12578,26 +27999,18 @@
   },
   {
     "id": "gpt-audio-1.5",
-    "name": "GPT-Audio 1.5",
+    "name": "gpt-audio-1.5",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2026-02-20 06:58:05 +0530",
+    "family": null,
+    "created_at": "2026-02-20 01:28:05 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -12613,26 +28026,18 @@
   },
   {
     "id": "gpt-audio-2025-08-28",
-    "name": "GPT-Audio 20250828",
+    "name": "gpt-audio-2025-08-28",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-08-27 06:25:46 +0530",
+    "family": null,
+    "created_at": "2025-08-27 00:55:46 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -12648,26 +28053,18 @@
   },
   {
     "id": "gpt-audio-mini",
-    "name": "GPT-Audio Mini",
+    "name": "gpt-audio-mini",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-10-03 22:50:27 +0530",
+    "family": null,
+    "created_at": "2025-10-03 17:20:27 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -12683,26 +28080,18 @@
   },
   {
     "id": "gpt-audio-mini-2025-10-06",
-    "name": "GPT-Audio Mini 20251006",
+    "name": "gpt-audio-mini-2025-10-06",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-10-03 22:52:17 +0530",
+    "family": null,
+    "created_at": "2025-10-03 17:22:17 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -12718,26 +28107,18 @@
   },
   {
     "id": "gpt-audio-mini-2025-12-15",
-    "name": "GPT-Audio Mini 20251215",
+    "name": "gpt-audio-mini-2025-12-15",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-12-15 06:23:28 +0530",
+    "family": null,
+    "created_at": "2025-12-15 00:53:28 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -12756,7 +28137,7 @@
     "name": "gpt-image-1",
     "provider": "openai",
     "family": "gpt-image",
-    "created_at": "2025-04-24 00:00:00 +0530",
+    "created_at": "2025-04-24 00:00:00 UTC",
     "context_window": 0,
     "max_output_tokens": 0,
     "knowledge_cutoff": null,
@@ -12770,10 +28151,23 @@
       ]
     },
     "capabilities": [
-      "vision",
-      "streaming"
+      "vision"
     ],
-    "pricing": {},
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5.0,
+          "cache_read_input_per_million": 1.25
+        }
+      },
+      "images": {
+        "standard": {
+          "input_per_million": 10.0,
+          "output_per_million": 40.0,
+          "cache_read_input_per_million": 2.5
+        }
+      }
+    },
     "metadata": {
       "object": "model",
       "owned_by": "system",
@@ -12795,7 +28189,7 @@
     "name": "gpt-image-1-mini",
     "provider": "openai",
     "family": "gpt-image",
-    "created_at": "2025-09-26 00:00:00 +0530",
+    "created_at": "2025-09-26 00:00:00 UTC",
     "context_window": 0,
     "max_output_tokens": 0,
     "knowledge_cutoff": null,
@@ -12810,10 +28204,23 @@
       ]
     },
     "capabilities": [
-      "vision",
-      "streaming"
+      "vision"
     ],
-    "pricing": {},
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "cache_read_input_per_million": 0.2
+        }
+      },
+      "images": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 8.0,
+          "cache_read_input_per_million": 0.25
+        }
+      }
+    },
     "metadata": {
       "object": "model",
       "owned_by": "system",
@@ -12835,7 +28242,7 @@
     "name": "gpt-image-1.5",
     "provider": "openai",
     "family": "gpt-image",
-    "created_at": "2025-11-25 00:00:00 +0530",
+    "created_at": "2025-11-25 00:00:00 UTC",
     "context_window": 0,
     "max_output_tokens": 0,
     "knowledge_cutoff": null,
@@ -12850,10 +28257,23 @@
       ]
     },
     "capabilities": [
-      "vision",
-      "streaming"
+      "vision"
     ],
-    "pricing": {},
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5.0,
+          "cache_read_input_per_million": 1.25
+        }
+      },
+      "images": {
+        "standard": {
+          "input_per_million": 8.0,
+          "output_per_million": 32.0,
+          "cache_read_input_per_million": 2.0
+        }
+      }
+    },
     "metadata": {
       "object": "model",
       "owned_by": "system",
@@ -12871,25 +28291,73 @@
     }
   },
   {
-    "id": "gpt-realtime",
-    "name": "GPT-Realtime",
+    "id": "gpt-image-2",
+    "name": "gpt-image-2",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-08-27 10:45:01 +0530",
+    "family": null,
+    "created_at": "2026-04-17 04:23:15 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": "system"
+    }
+  },
+  {
+    "id": "gpt-image-2-2026-04-21",
+    "name": "gpt-image-2-2026-04-21",
+    "provider": "openai",
+    "family": null,
+    "created_at": "2026-04-17 04:26:34 UTC",
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "object": "model",
+      "owned_by": "system"
+    }
+  },
+  {
+    "id": "gpt-realtime",
+    "name": "gpt-realtime",
+    "provider": "openai",
+    "family": null,
+    "created_at": "2025-08-27 05:15:01 UTC",
+    "context_window": 4096,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -12905,24 +28373,18 @@
   },
   {
     "id": "gpt-realtime-1.5",
-    "name": "GPT-Realtime 1.5",
+    "name": "gpt-realtime-1.5",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2026-02-19 06:07:49 +0530",
+    "family": null,
+    "created_at": "2026-02-19 00:37:49 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -12938,24 +28400,18 @@
   },
   {
     "id": "gpt-realtime-2025-08-28",
-    "name": "GPT-Realtime 20250828",
+    "name": "gpt-realtime-2025-08-28",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-08-27 10:46:13 +0530",
+    "family": null,
+    "created_at": "2025-08-27 05:16:13 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -12971,24 +28427,18 @@
   },
   {
     "id": "gpt-realtime-mini",
-    "name": "GPT-Realtime Mini",
+    "name": "gpt-realtime-mini",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-10-04 00:15:33 +0530",
+    "family": null,
+    "created_at": "2025-10-03 18:45:33 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -13004,24 +28454,18 @@
   },
   {
     "id": "gpt-realtime-mini-2025-10-06",
-    "name": "GPT-Realtime Mini 20251006",
+    "name": "gpt-realtime-mini-2025-10-06",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-10-04 00:16:15 +0530",
+    "family": null,
+    "created_at": "2025-10-03 18:46:15 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -13037,24 +28481,18 @@
   },
   {
     "id": "gpt-realtime-mini-2025-12-15",
-    "name": "GPT-Realtime Mini 20251215",
+    "name": "gpt-realtime-mini-2025-12-15",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-12-13 13:16:47 +0530",
+    "family": null,
+    "created_at": "2025-12-13 07:46:47 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -13073,7 +28511,7 @@
     "name": "o1",
     "provider": "openai",
     "family": "o",
-    "created_at": "2024-12-05 00:00:00 +0530",
+    "created_at": "2024-12-05 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 100000,
     "knowledge_cutoff": null,
@@ -13091,15 +28529,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 15,
           "output_per_million": 60,
-          "cached_input_per_million": 7.5
+          "cache_read_input_per_million": 7.5
         }
       }
     },
@@ -13126,27 +28563,21 @@
   },
   {
     "id": "o1-2024-12-17",
-    "name": "O1-20241217",
+    "name": "o1-2024-12-17",
     "provider": "openai",
-    "family": "o1",
-    "created_at": "2024-12-16 10:59:36 +0530",
+    "family": null,
+    "created_at": "2024-12-16 05:29:36 UTC",
     "context_window": 200000,
     "max_output_tokens": 100000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
       "structured_output",
+      "vision",
       "reasoning"
     ],
     "pricing": {
@@ -13167,7 +28598,7 @@
     "name": "o1-mini",
     "provider": "openai",
     "family": "o-mini",
-    "created_at": "2024-09-12 00:00:00 +0530",
+    "created_at": "2024-09-12 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -13188,7 +28619,7 @@
         "standard": {
           "input_per_million": 1.1,
           "output_per_million": 4.4,
-          "cached_input_per_million": 0.55
+          "cache_read_input_per_million": 0.55
         }
       }
     },
@@ -13216,7 +28647,7 @@
     "name": "o1-preview",
     "provider": "openai",
     "family": "o",
-    "created_at": "2024-09-12 00:00:00 +0530",
+    "created_at": "2024-09-12 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -13236,7 +28667,7 @@
         "standard": {
           "input_per_million": 15,
           "output_per_million": 60,
-          "cached_input_per_million": 7.5
+          "cache_read_input_per_million": 7.5
         }
       }
     },
@@ -13264,7 +28695,7 @@
     "name": "o1-pro",
     "provider": "openai",
     "family": "o-pro",
-    "created_at": "2025-03-19 00:00:00 +0530",
+    "created_at": "2025-03-19 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 100000,
     "knowledge_cutoff": null,
@@ -13281,8 +28712,7 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
@@ -13314,27 +28744,21 @@
   },
   {
     "id": "o1-pro-2025-03-19",
-    "name": "O1-Pro 20250319",
+    "name": "o1-pro-2025-03-19",
     "provider": "openai",
-    "family": "o1_pro",
-    "created_at": "2025-03-18 04:15:04 +0530",
+    "family": null,
+    "created_at": "2025-03-17 22:45:04 UTC",
     "context_window": 200000,
     "max_output_tokens": 100000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
       "structured_output",
+      "vision",
       "reasoning"
     ],
     "pricing": {
@@ -13355,7 +28779,7 @@
     "name": "o3",
     "provider": "openai",
     "family": "o",
-    "created_at": "2025-04-16 00:00:00 +0530",
+    "created_at": "2025-04-16 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 100000,
     "knowledge_cutoff": null,
@@ -13373,15 +28797,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 2,
           "output_per_million": 8,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5
         }
       }
     },
@@ -13408,23 +28831,18 @@
   },
   {
     "id": "o3-2025-04-16",
-    "name": "O3-20250416",
+    "name": "o3-2025-04-16",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-04-08 22:58:21 +0530",
+    "family": null,
+    "created_at": "2025-04-08 17:28:21 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "reasoning"
     ],
     "pricing": {
@@ -13445,7 +28863,7 @@
     "name": "o3-deep-research",
     "provider": "openai",
     "family": "o",
-    "created_at": "2024-06-26 00:00:00 +0530",
+    "created_at": "2024-06-26 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 100000,
     "knowledge_cutoff": null,
@@ -13461,15 +28879,14 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 10,
           "output_per_million": 40,
-          "cached_input_per_million": 2.5
+          "cache_read_input_per_million": 2.5
         }
       }
     },
@@ -13496,23 +28913,18 @@
   },
   {
     "id": "o3-deep-research-2025-06-26",
-    "name": "O3-Deep Research 20250626",
+    "name": "o3-deep-research-2025-06-26",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-06-25 20:56:59 +0530",
+    "family": null,
+    "created_at": "2025-06-25 15:26:59 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "reasoning"
     ],
     "pricing": {
@@ -13533,7 +28945,7 @@
     "name": "o3-mini",
     "provider": "openai",
     "family": "o-mini",
-    "created_at": "2024-12-20 00:00:00 +0530",
+    "created_at": "2024-12-20 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 100000,
     "knowledge_cutoff": null,
@@ -13548,15 +28960,14 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning",
-      "streaming"
+      "reasoning"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.1,
           "output_per_million": 4.4,
-          "cached_input_per_million": 0.55
+          "cache_read_input_per_million": 0.55
         }
       }
     },
@@ -13583,23 +28994,18 @@
   },
   {
     "id": "o3-mini-2025-01-31",
-    "name": "O3-Mini 20250131",
+    "name": "o3-mini-2025-01-31",
     "provider": "openai",
-    "family": "o3_mini",
-    "created_at": "2025-01-28 02:06:40 +0530",
+    "family": null,
+    "created_at": "2025-01-27 20:36:40 UTC",
     "context_window": 200000,
     "max_output_tokens": 100000,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "function_calling",
       "structured_output",
       "reasoning"
@@ -13622,7 +29028,7 @@
     "name": "o3-pro",
     "provider": "openai",
     "family": "o-pro",
-    "created_at": "2025-06-10 00:00:00 +0530",
+    "created_at": "2025-06-10 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 100000,
     "knowledge_cutoff": null,
@@ -13639,8 +29045,7 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
@@ -13672,23 +29077,18 @@
   },
   {
     "id": "o3-pro-2025-06-10",
-    "name": "O3-Pro 20250610",
+    "name": "o3-pro-2025-06-10",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-06-06 05:09:21 +0530",
+    "family": null,
+    "created_at": "2025-06-05 23:39:21 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "reasoning"
     ],
     "pricing": {
@@ -13709,7 +29109,7 @@
     "name": "o4-mini",
     "provider": "openai",
     "family": "o-mini",
-    "created_at": "2025-04-16 00:00:00 +0530",
+    "created_at": "2025-04-16 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 100000,
     "knowledge_cutoff": null,
@@ -13726,15 +29126,14 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.1,
           "output_per_million": 4.4,
-          "cached_input_per_million": 0.28
+          "cache_read_input_per_million": 0.28
         }
       }
     },
@@ -13761,23 +29160,18 @@
   },
   {
     "id": "o4-mini-2025-04-16",
-    "name": "O4 Mini 20250416",
+    "name": "o4-mini-2025-04-16",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-04-08 23:01:46 +0530",
+    "family": null,
+    "created_at": "2025-04-08 17:31:46 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "reasoning"
     ],
     "pricing": {
@@ -13798,7 +29192,7 @@
     "name": "o4-mini-deep-research",
     "provider": "openai",
     "family": "o-mini",
-    "created_at": "2024-06-26 00:00:00 +0530",
+    "created_at": "2024-06-26 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 100000,
     "knowledge_cutoff": null,
@@ -13814,15 +29208,14 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision",
-      "streaming"
+      "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 2,
           "output_per_million": 8,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5
         }
       }
     },
@@ -13849,23 +29242,18 @@
   },
   {
     "id": "o4-mini-deep-research-2025-06-26",
-    "name": "O4 Mini Deep Research 20250626",
+    "name": "o4-mini-deep-research-2025-06-26",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-06-25 21:12:01 +0530",
+    "family": null,
+    "created_at": "2025-06-25 15:42:01 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
     "capabilities": [
-      "streaming",
       "reasoning"
     ],
     "pricing": {
@@ -13883,25 +29271,20 @@
   },
   {
     "id": "omni-moderation-2024-09-26",
-    "name": "Omni Moderation 20240926",
+    "name": "omni-moderation-2024-09-26",
     "provider": "openai",
-    "family": "moderation",
-    "created_at": "2024-11-28 00:37:46 +0530",
+    "family": null,
+    "created_at": "2024-11-27 19:07:46 UTC",
     "context_window": null,
     "max_output_tokens": null,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text",
-        "moderation"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [],
+    "capabilities": [
+      "vision"
+    ],
     "pricing": {},
     "metadata": {
       "object": "model",
@@ -13910,25 +29293,20 @@
   },
   {
     "id": "omni-moderation-latest",
-    "name": "Omni Moderation Latest",
+    "name": "omni-moderation-latest",
     "provider": "openai",
-    "family": "moderation",
-    "created_at": "2024-11-15 22:17:45 +0530",
+    "family": null,
+    "created_at": "2024-11-15 16:47:45 UTC",
     "context_window": null,
     "max_output_tokens": null,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "image",
-        "pdf"
-      ],
-      "output": [
-        "text",
-        "moderation"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [],
+    "capabilities": [
+      "vision"
+    ],
     "pricing": {},
     "metadata": {
       "object": "model",
@@ -13937,24 +29315,18 @@
   },
   {
     "id": "sora-2",
-    "name": "Sora 2",
+    "name": "sora-2",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-10-06 05:26:55 +0530",
+    "family": null,
+    "created_at": "2025-10-05 23:56:55 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -13970,24 +29342,18 @@
   },
   {
     "id": "sora-2-pro",
-    "name": "Sora 2 Pro",
+    "name": "sora-2-pro",
     "provider": "openai",
-    "family": "other",
-    "created_at": "2025-10-06 05:27:43 +0530",
+    "family": null,
+    "created_at": "2025-10-05 23:57:43 UTC",
     "context_window": 4096,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -14006,7 +29372,7 @@
     "name": "text-embedding-3-large",
     "provider": "openai",
     "family": "text-embedding",
-    "created_at": "2024-01-25 00:00:00 +0530",
+    "created_at": "2024-01-25 00:00:00 UTC",
     "context_window": 8191,
     "max_output_tokens": 3072,
     "knowledge_cutoff": null,
@@ -14015,12 +29381,10 @@
         "text"
       ],
       "output": [
-        "text"
+        "embeddings"
       ]
     },
-    "capabilities": [
-      "batch"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -14053,7 +29417,7 @@
     "name": "text-embedding-3-small",
     "provider": "openai",
     "family": "text-embedding",
-    "created_at": "2024-01-25 00:00:00 +0530",
+    "created_at": "2024-01-25 00:00:00 UTC",
     "context_window": 8191,
     "max_output_tokens": 1536,
     "knowledge_cutoff": null,
@@ -14062,12 +29426,10 @@
         "text"
       ],
       "output": [
-        "text"
+        "embeddings"
       ]
     },
-    "capabilities": [
-      "batch"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -14100,7 +29462,7 @@
     "name": "text-embedding-ada-002",
     "provider": "openai",
     "family": "text-embedding",
-    "created_at": "2022-12-15 00:00:00 +0530",
+    "created_at": "2022-12-15 00:00:00 UTC",
     "context_window": 8192,
     "max_output_tokens": 1536,
     "knowledge_cutoff": null,
@@ -14109,12 +29471,10 @@
         "text"
       ],
       "output": [
-        "text"
+        "embeddings"
       ]
     },
-    "capabilities": [
-      "batch"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -14144,26 +29504,18 @@
   },
   {
     "id": "tts-1",
-    "name": "TTS-1",
+    "name": "tts-1",
     "provider": "openai",
-    "family": "tts1",
-    "created_at": "2023-04-20 03:19:11 +0530",
+    "family": null,
+    "created_at": "2023-04-19 21:49:11 UTC",
     "context_window": null,
     "max_output_tokens": null,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -14179,26 +29531,18 @@
   },
   {
     "id": "tts-1-1106",
-    "name": "TTS-1 1106",
+    "name": "tts-1-1106",
     "provider": "openai",
-    "family": "tts1",
-    "created_at": "2023-11-04 04:44:01 +0530",
+    "family": null,
+    "created_at": "2023-11-03 23:14:01 UTC",
     "context_window": null,
     "max_output_tokens": null,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -14214,26 +29558,18 @@
   },
   {
     "id": "tts-1-hd",
-    "name": "TTS-1 HD",
+    "name": "tts-1-hd",
     "provider": "openai",
-    "family": "tts1_hd",
-    "created_at": "2023-11-04 02:43:35 +0530",
+    "family": null,
+    "created_at": "2023-11-03 21:13:35 UTC",
     "context_window": null,
     "max_output_tokens": null,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -14249,26 +29585,18 @@
   },
   {
     "id": "tts-1-hd-1106",
-    "name": "TTS-1 HD 1106",
+    "name": "tts-1-hd-1106",
     "provider": "openai",
-    "family": "tts1_hd",
-    "created_at": "2023-11-04 04:48:53 +0530",
+    "family": null,
+    "created_at": "2023-11-03 23:18:53 UTC",
     "context_window": null,
     "max_output_tokens": null,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text",
-        "audio"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -14284,25 +29612,18 @@
   },
   {
     "id": "whisper-1",
-    "name": "Whisper 1",
+    "name": "whisper-1",
     "provider": "openai",
-    "family": "whisper",
-    "created_at": "2023-02-28 02:43:04 +0530",
+    "family": null,
+    "created_at": "2023-02-27 21:13:04 UTC",
     "context_window": null,
     "max_output_tokens": null,
     "knowledge_cutoff": null,
     "modalities": {
-      "input": [
-        "text",
-        "audio"
-      ],
-      "output": [
-        "text"
-      ]
+      "input": [],
+      "output": []
     },
-    "capabilities": [
-      "streaming"
-    ],
+    "capabilities": [],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -14316,12 +29637,1068 @@
       "owned_by": "openai-internal"
     }
   },
+  {
+    "id": "ai21/jamba-large-1.7",
+    "name": "AI21: Jamba Large 1.7",
+    "provider": "openrouter",
+    "family": "ai21",
+    "created_at": "2025-08-08 16:03:40 UTC",
+    "context_window": 256000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "output_per_million": 8.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "Jamba Large 1.7 is the latest model in the Jamba open family, offering improvements in grounding, instruction-following, and overall efficiency. Built on a hybrid SSM-Transformer architecture with a 256K context...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 256000,
+        "max_completion_tokens": 4096,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "response_format",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "aion-labs/aion-1.0",
+    "name": "AionLabs: Aion-1.0",
+    "provider": "openrouter",
+    "family": "aion-labs",
+    "created_at": "2025-02-04 19:32:37 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 4.0,
+          "output_per_million": 8.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "Aion-1.0 is a multi-model system designed for high performance across various tasks, including reasoning and coding. It is built on DeepSeek-R1, augmented with additional models and techniques such as Tree...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "temperature",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "aion-labs/aion-1.0-mini",
+    "name": "AionLabs: Aion-1.0-Mini",
+    "provider": "openrouter",
+    "family": "aion-labs",
+    "created_at": "2025-02-04 19:25:07 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.7,
+          "output_per_million": 1.4
+        }
+      }
+    },
+    "metadata": {
+      "description": "Aion-1.0-Mini 32B parameter model is a distilled version of the DeepSeek-R1 model, designed for strong performance in reasoning domains such as mathematics, coding, and logic. It is a modified variant...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "temperature",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "aion-labs/aion-2.0",
+    "name": "AionLabs: Aion-2.0",
+    "provider": "openrouter",
+    "family": "aion-labs",
+    "created_at": "2026-02-23 21:15:06 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.7999999999999999,
+          "output_per_million": 1.5999999999999999,
+          "cache_read_input_per_million": 0.19999999999999998
+        }
+      }
+    },
+    "metadata": {
+      "description": "Aion-2.0 is a variant of DeepSeek V3.2 optimized for immersive roleplaying and storytelling. It is particularly strong at introducing tension, crises, and conflict into stories, making narratives feel more engaging....",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "temperature",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "aion-labs/aion-rp-llama-3.1-8b",
+    "name": "AionLabs: Aion-RP 1.0 (8B)",
+    "provider": "openrouter",
+    "family": "aion-labs",
+    "created_at": "2025-02-04 19:18:38 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.7999999999999999,
+          "output_per_million": 1.5999999999999999
+        }
+      }
+    },
+    "metadata": {
+      "description": "Aion-RP-Llama-3.1-8B ranks the highest in the character evaluation portion of the RPBench-Auto benchmark, a roleplaying-specific variant of Arena-Hard-Auto, where LLMs evaluate each other’s responses. It is a fine-tuned base model...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "temperature",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "alfredpros/codellama-7b-instruct-solidity",
+    "name": "AlfredPros: CodeLLaMa 7B Instruct Solidity",
+    "provider": "openrouter",
+    "family": "alfredpros",
+    "created_at": "2025-04-14 14:44:34 UTC",
+    "context_window": 4096,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.7999999999999999,
+          "output_per_million": 1.2
+        }
+      }
+    },
+    "metadata": {
+      "description": "A finetuned 7 billion parameters Code LLaMA - Instruct model to generate Solidity smart contract using 4-bit QLoRA finetuning provided by PEFT library.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": "alpaca"
+      },
+      "top_provider": {
+        "context_length": 4096,
+        "max_completion_tokens": 4096,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "alibaba/tongyi-deepresearch-30b-a3b",
+    "name": "Tongyi DeepResearch 30B A3B",
+    "provider": "openrouter",
+    "family": "alibaba",
+    "created_at": "2025-09-18 15:53:24 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 131072,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.09,
+          "output_per_million": 0.44999999999999996,
+          "cache_read_input_per_million": 0.09
+        }
+      }
+    },
+    "metadata": {
+      "description": "Tongyi DeepResearch is an agentic large language model developed by Tongyi Lab, with 30 billion total parameters activating only 3 billion per token. It's optimized for long-horizon, deep information-seeking tasks...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 131072,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "allenai/olmo-3-32b-think",
+    "name": "AllenAI: Olmo 3 32B Think",
+    "provider": "openrouter",
+    "family": "allenai",
+    "created_at": "2025-11-21 20:51:16 UTC",
+    "context_window": 65536,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.15,
+          "output_per_million": 0.5
+        }
+      }
+    },
+    "metadata": {
+      "description": "Olmo 3 32B Think is a large-scale, 32-billion-parameter model purpose-built for deep reasoning, complex logic chains and advanced instruction-following scenarios. Its capacity enables strong performance on demanding evaluation tasks and...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 65536,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "allenai/olmo-3.1-32b-instruct",
+    "name": "AllenAI: Olmo 3.1 32B Instruct",
+    "provider": "openrouter",
+    "family": "allenai",
+    "created_at": "2026-01-06 19:42:34 UTC",
+    "context_window": 65536,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.19999999999999998,
+          "output_per_million": 0.6
+        }
+      }
+    },
+    "metadata": {
+      "description": "Olmo 3.1 32B Instruct is a large-scale, 32-billion-parameter instruction-tuned language model engineered for high-performance conversational AI, multi-turn dialogue, and practical instruction following. As part of the Olmo 3.1 family, this...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 65536,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "alpindale/goliath-120b",
+    "name": "Goliath 120B",
+    "provider": "openrouter",
+    "family": "alpindale",
+    "created_at": "2023-11-10 00:00:00 UTC",
+    "context_window": 6144,
+    "max_output_tokens": 1024,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 3.75,
+          "output_per_million": 7.5
+        }
+      }
+    },
+    "metadata": {
+      "description": "A large LLM created by combining two fine-tuned Llama 70B models into one 120B model. Combines Xwin and Euryale. Credits to - [@chargoddard](https://huggingface.co/chargoddard) for developing the framework used to merge...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama2",
+        "instruct_type": "airoboros"
+      },
+      "top_provider": {
+        "context_length": 6144,
+        "max_completion_tokens": 1024,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "top_a",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "amazon/nova-2-lite-v1",
+    "name": "Amazon: Nova 2 Lite",
+    "provider": "openrouter",
+    "family": "amazon",
+    "created_at": "2025-12-02 17:31:12 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 65535,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.3,
+          "output_per_million": 2.5
+        }
+      }
+    },
+    "metadata": {
+      "description": "Nova 2 Lite is a fast, cost-effective reasoning model for everyday workloads that can process text, images, and videos to generate text. Nova 2 Lite demonstrates standout capabilities in processing...",
+      "architecture": {
+        "modality": "text+image+file+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "video",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Nova",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 65535,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "amazon/nova-lite-v1",
+    "name": "Amazon: Nova Lite 1.0",
+    "provider": "openrouter",
+    "family": "amazon",
+    "created_at": "2024-12-05 22:22:43 UTC",
+    "context_window": 300000,
+    "max_output_tokens": 5120,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.06,
+          "output_per_million": 0.24
+        }
+      }
+    },
+    "metadata": {
+      "description": "Amazon Nova Lite 1.0 is a very low-cost multimodal model from Amazon that focused on fast processing of image, video, and text inputs to generate text output. Amazon Nova Lite...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Nova",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 300000,
+        "max_completion_tokens": 5120,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "stop",
+        "temperature",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "amazon/nova-micro-v1",
+    "name": "Amazon: Nova Micro 1.0",
+    "provider": "openrouter",
+    "family": "amazon",
+    "created_at": "2024-12-05 22:20:37 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 5120,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.035,
+          "output_per_million": 0.14
+        }
+      }
+    },
+    "metadata": {
+      "description": "Amazon Nova Micro 1.0 is a text-only model that delivers the lowest latency responses in the Amazon Nova family of models at a very low cost. With a context length...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Nova",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 5120,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "stop",
+        "temperature",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "amazon/nova-premier-v1",
+    "name": "Amazon: Nova Premier 1.0",
+    "provider": "openrouter",
+    "family": "amazon",
+    "created_at": "2025-10-31 22:38:52 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 32000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 12.5,
+          "cache_read_input_per_million": 0.625
+        }
+      }
+    },
+    "metadata": {
+      "description": "Amazon Nova Premier is the most capable of Amazon’s multimodal models for complex reasoning tasks and for use as the best teacher for distilling custom models.",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Nova",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 32000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "stop",
+        "temperature",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "amazon/nova-pro-v1",
+    "name": "Amazon: Nova Pro 1.0",
+    "provider": "openrouter",
+    "family": "amazon",
+    "created_at": "2024-12-05 22:05:03 UTC",
+    "context_window": 300000,
+    "max_output_tokens": 5120,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.7999999999999999,
+          "output_per_million": 3.1999999999999997
+        }
+      }
+    },
+    "metadata": {
+      "description": "Amazon Nova Pro 1.0 is a capable multimodal model from Amazon focused on providing a combination of accuracy, speed, and cost for a wide range of tasks. As of December...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Nova",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 300000,
+        "max_completion_tokens": 5120,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "stop",
+        "temperature",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "anthracite-org/magnum-v4-72b",
+    "name": "Magnum v4 72B",
+    "provider": "openrouter",
+    "family": "anthracite-org",
+    "created_at": "2024-10-22 00:00:00 UTC",
+    "context_window": 16384,
+    "max_output_tokens": 2048,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 3.0,
+          "output_per_million": 5.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "This is a series of models designed to replicate the prose quality of the Claude 3 models, specifically Sonnet(https://openrouter.ai/anthropic/claude-3.5-sonnet) and Opus(https://openrouter.ai/anthropic/claude-3-opus).\n\nThe model is fine-tuned on top of [Qwen2.5 72B](https://openrouter.ai/qwen/qwen-2.5-72b-instruct).",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen",
+        "instruct_type": "chatml"
+      },
+      "top_provider": {
+        "context_length": 16384,
+        "max_completion_tokens": 2048,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "top_a",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "anthropic/claude-3-haiku",
+    "name": "Anthropic: Claude 3 Haiku",
+    "provider": "openrouter",
+    "family": "anthropic",
+    "created_at": "2024-03-13 00:00:00 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.25,
+          "output_per_million": 1.25,
+          "cache_read_input_per_million": 0.03
+        }
+      }
+    },
+    "metadata": {
+      "description": "Claude 3 Haiku is Anthropic's fastest and most compact model for\nnear-instant responsiveness. Quick and accurate targeted performance.\n\nSee the launch announcement and benchmark results [here](https://www.anthropic.com/news/claude-3-haiku)\n\n#multimodal",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Claude",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 4096,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "anthropic/claude-3.5-haiku",
     "name": "Claude Haiku 3.5",
     "provider": "openrouter",
     "family": "claude-haiku",
-    "created_at": "2024-10-22 00:00:00 +0530",
+    "created_at": "2024-10-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": "2024-07-31",
@@ -14337,18 +30714,48 @@
     },
     "capabilities": [
       "function_calling",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.8,
           "output_per_million": 4,
-          "cached_input_per_million": 0.08
+          "cache_read_input_per_million": 0.08,
+          "cache_write_input_per_million": 1
         }
       }
     },
     "metadata": {
+      "description": "Claude 3.5 Haiku features offers enhanced capabilities in speed, coding accuracy, and tool use. Engineered to excel in real-time applications, it delivers quick response times that are essential for dynamic...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Claude",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 8192,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -14373,7 +30780,7 @@
     "name": "Claude Sonnet 3.7",
     "provider": "openrouter",
     "family": "claude-sonnet",
-    "created_at": "2025-02-19 00:00:00 +0530",
+    "created_at": "2025-02-19 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": null,
@@ -14390,18 +30797,50 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 15,
           "output_per_million": 75,
-          "cached_input_per_million": 1.5
+          "cache_read_input_per_million": 1.5,
+          "cache_write_input_per_million": 18.75
         }
       }
     },
     "metadata": {
+      "description": "Claude 3.7 Sonnet is an advanced large language model with improved reasoning, coding, and problem-solving capabilities. It introduces a hybrid reasoning approach, allowing users to choose between rapid responses and...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Claude",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 64000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -14421,12 +30860,77 @@
       "knowledge": "2024-01"
     }
   },
+  {
+    "id": "anthropic/claude-3.7-sonnet:thinking",
+    "name": "Anthropic: Claude 3.7 Sonnet (thinking)",
+    "provider": "openrouter",
+    "family": "anthropic",
+    "created_at": "2025-02-24 18:35:10 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 64000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 3.0,
+          "output_per_million": 15.0,
+          "cache_read_input_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "description": "Claude 3.7 Sonnet is an advanced large language model with improved reasoning, coding, and problem-solving capabilities. It introduces a hybrid reasoning approach, allowing users to choose between rapid responses and...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Claude",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 64000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "anthropic/claude-haiku-4.5",
     "name": "Claude Haiku 4.5",
     "provider": "openrouter",
     "family": "claude-haiku",
-    "created_at": "2025-10-15 00:00:00 +0530",
+    "created_at": "2025-10-15 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-02-28",
@@ -14444,18 +30948,52 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1,
           "output_per_million": 5,
-          "cached_input_per_million": 0.1
+          "cache_read_input_per_million": 0.1,
+          "cache_write_input_per_million": 1.25
         }
       }
     },
     "metadata": {
+      "description": "Claude Haiku 4.5 is Anthropic’s fastest and most efficient model, delivering near-frontier intelligence at a fraction of the cost and latency of larger Claude models. Matching Claude Sonnet 4’s performance...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Claude",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 64000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -14480,7 +31018,7 @@
     "name": "Claude Opus 4",
     "provider": "openrouter",
     "family": "claude-opus",
-    "created_at": "2025-05-22 00:00:00 +0530",
+    "created_at": "2025-05-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 32000,
     "knowledge_cutoff": "2025-03-31",
@@ -14497,18 +31035,51 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 15,
           "output_per_million": 75,
-          "cached_input_per_million": 1.5
+          "cache_read_input_per_million": 1.5,
+          "cache_write_input_per_million": 18.75
         }
       }
     },
     "metadata": {
+      "description": "Claude Opus 4 is benchmarked as the world’s best coding model, at time of release, bringing sustained performance on complex, long-running tasks and agent workflows. It sets new benchmarks in...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Claude",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 32000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -14533,7 +31104,7 @@
     "name": "Claude Opus 4.1",
     "provider": "openrouter",
     "family": "claude-opus",
-    "created_at": "2025-08-05 00:00:00 +0530",
+    "created_at": "2025-08-05 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 32000,
     "knowledge_cutoff": "2025-03-31",
@@ -14551,18 +31122,53 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 15,
           "output_per_million": 75,
-          "cached_input_per_million": 1.5
+          "cache_read_input_per_million": 1.5,
+          "cache_write_input_per_million": 18.75
         }
       }
     },
     "metadata": {
+      "description": "Claude Opus 4.1 is an updated version of Anthropic’s flagship model, offering improved performance in coding, reasoning, and agentic tasks. It achieves 74.5% on SWE-bench Verified and shows notable gains...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Claude",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 32000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -14587,7 +31193,7 @@
     "name": "Claude Opus 4.5",
     "provider": "openrouter",
     "family": "claude-opus",
-    "created_at": "2025-11-24 00:00:00 +0530",
+    "created_at": "2025-11-24 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 32000,
     "knowledge_cutoff": "2025-05-30",
@@ -14605,18 +31211,53 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 5,
           "output_per_million": 25,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
         }
       }
     },
     "metadata": {
+      "description": "Claude Opus 4.5 is Anthropic’s frontier reasoning model optimized for complex software engineering, agentic workflows, and long-horizon computer use. It offers strong multimodal capabilities, competitive performance across real-world coding and...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "file",
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Claude",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 64000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "verbosity"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -14641,10 +31282,10 @@
     "name": "Claude Opus 4.6",
     "provider": "openrouter",
     "family": "claude-opus",
-    "created_at": "2026-02-05 00:00:00 +0530",
+    "created_at": "2026-02-05 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 128000,
-    "knowledge_cutoff": "2025-05-30",
+    "knowledge_cutoff": "2025-05-31",
     "modalities": {
       "input": [
         "text",
@@ -14659,18 +31300,54 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 5,
           "output_per_million": 25,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
         }
       }
     },
     "metadata": {
+      "description": "Opus 4.6 is Anthropic’s strongest model for coding and long-running professional tasks. It is built for agents that operate across entire workflows rather than single prompts, making it especially effective...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Claude",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 128000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p",
+        "verbosity"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -14693,7 +31370,166 @@
         "context": 1000000,
         "output": 128000
       },
-      "knowledge": "2025-05-30"
+      "knowledge": "2025-05-31"
+    }
+  },
+  {
+    "id": "anthropic/claude-opus-4.6-fast",
+    "name": "Anthropic: Claude Opus 4.6 (Fast)",
+    "provider": "openrouter",
+    "family": "anthropic",
+    "created_at": "2026-04-07 20:07:52 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 30.0,
+          "output_per_million": 150.0,
+          "cache_read_input_per_million": 3.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "Fast-mode variant of [Opus 4.6](/anthropic/claude-opus-4.6) - identical capabilities with higher output speed at premium 6x pricing.\n\nLearn more in Anthropic's docs: https://platform.claude.com/docs/en/build-with-claude/fast-mode",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Claude",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 128000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p",
+        "verbosity"
+      ]
+    }
+  },
+  {
+    "id": "anthropic/claude-opus-4.7",
+    "name": "Claude Opus 4.7",
+    "provider": "openrouter",
+    "family": "claude-opus",
+    "created_at": "2026-04-16 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": "2026-01-31",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5,
+          "output_per_million": 25,
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
+        }
+      }
+    },
+    "metadata": {
+      "description": "Opus 4.7 is the next generation of Anthropic's Opus family, built for long-running, asynchronous agents. Building on the coding and agentic strengths of Opus 4.6, it delivers stronger performance on...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Claude",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 128000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "tool_choice",
+        "tools",
+        "verbosity"
+      ],
+      "source": "models.dev",
+      "provider_id": "openrouter",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": false,
+      "last_updated": "2026-04-16",
+      "cost": {
+        "input": 5,
+        "output": 25,
+        "cache_read": 0.5,
+        "cache_write": 6.25,
+        "context_over_200k": {
+          "input": 10,
+          "output": 37.5,
+          "cache_read": 1,
+          "cache_write": 12.5
+        }
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 128000
+      },
+      "knowledge": "2026-01-31"
     }
   },
   {
@@ -14701,7 +31537,7 @@
     "name": "Claude Sonnet 4",
     "provider": "openrouter",
     "family": "claude-sonnet",
-    "created_at": "2025-05-22 00:00:00 +0530",
+    "created_at": "2025-05-22 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-03-31",
@@ -14718,18 +31554,51 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
     "metadata": {
+      "description": "Claude Sonnet 4 significantly enhances the capabilities of its predecessor, Sonnet 3.7, excelling in both coding and reasoning tasks with improved precision and controllability. Achieving state-of-the-art performance on SWE-bench (72.7%),...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Claude",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 64000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -14760,7 +31629,7 @@
     "name": "Claude Sonnet 4.5",
     "provider": "openrouter",
     "family": "claude-sonnet",
-    "created_at": "2025-09-29 00:00:00 +0530",
+    "created_at": "2025-09-29 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": "2025-07-31",
@@ -14778,18 +31647,53 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
     "metadata": {
+      "description": "Claude Sonnet 4.5 is Anthropic’s most advanced Sonnet model to date, optimized for real-world agents and coding workflows. It delivers state-of-the-art performance on coding benchmarks such as SWE-bench Verified, with...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Claude",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 64000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -14820,10 +31724,10 @@
     "name": "Claude Sonnet 4.6",
     "provider": "openrouter",
     "family": "claude-sonnet",
-    "created_at": "2026-02-17 00:00:00 +0530",
+    "created_at": "2026-02-17 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 128000,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2025-08-31",
     "modalities": {
       "input": [
         "text",
@@ -14837,18 +31741,54 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.3
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
     "metadata": {
+      "description": "Sonnet 4.6 is Anthropic's most capable Sonnet-class model yet, with frontier performance across coding, agents, and professional work. It excels at iterative development, complex codebase navigation, end-to-end project management with...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Claude",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 128000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p",
+        "verbosity"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -14870,7 +31810,256 @@
       "limit": {
         "context": 1000000,
         "output": 128000
+      },
+      "knowledge": "2025-08-31"
+    }
+  },
+  {
+    "id": "arcee-ai/coder-large",
+    "name": "Arcee AI: Coder Large",
+    "provider": "openrouter",
+    "family": "arcee-ai",
+    "created_at": "2025-05-05 20:57:43 UTC",
+    "context_window": 32768,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 0.7999999999999999
+        }
       }
+    },
+    "metadata": {
+      "description": "Coder‑Large is a 32 B‑parameter offspring of Qwen 2.5‑Instruct that has been further trained on permissively‑licensed GitHub, CodeSearchNet and synthetic bug‑fix corpora. It supports a 32k context window, enabling multi‑file...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "arcee-ai/maestro-reasoning",
+    "name": "Arcee AI: Maestro Reasoning",
+    "provider": "openrouter",
+    "family": "arcee-ai",
+    "created_at": "2025-05-05 21:41:09 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 32000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.8999999999999999,
+          "output_per_million": 3.3000000000000003
+        }
+      }
+    },
+    "metadata": {
+      "description": "Maestro Reasoning is Arcee's flagship analysis model: a 32 B‑parameter derivative of Qwen 2.5‑32 B tuned with DPO and chain‑of‑thought RL for step‑by‑step logic. Compared to the earlier 7 B...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 32000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "arcee-ai/spotlight",
+    "name": "Arcee AI: Spotlight",
+    "provider": "openrouter",
+    "family": "arcee-ai",
+    "created_at": "2025-05-05 21:45:52 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 65537,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.18,
+          "output_per_million": 0.18
+        }
+      }
+    },
+    "metadata": {
+      "description": "Spotlight is a 7‑billion‑parameter vision‑language model derived from Qwen 2.5‑VL and fine‑tuned by Arcee AI for tight image‑text grounding tasks. It offers a 32 k‑token context window, enabling rich multimodal...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 65537,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "arcee-ai/trinity-large-preview",
+    "name": "Arcee AI: Trinity Large Preview",
+    "provider": "openrouter",
+    "family": "arcee-ai",
+    "created_at": "2026-01-27 22:24:30 UTC",
+    "context_window": 131000,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.15,
+          "output_per_million": 0.44999999999999996
+        }
+      }
+    },
+    "metadata": {
+      "description": "Trinity-Large-Preview is a frontier-scale open-weight language model from Arcee, built as a 400B-parameter sparse Mixture-of-Experts with 13B active parameters per token using 4-of-256 expert routing. It excels in creative writing,...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "response_format",
+        "structured_outputs",
+        "temperature",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
     }
   },
   {
@@ -14878,7 +32067,7 @@
     "name": "Trinity Large Preview",
     "provider": "openrouter",
     "family": "trinity",
-    "created_at": "2026-01-28 00:00:00 +0530",
+    "created_at": "2026-01-28 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -14918,7 +32107,7 @@
     "name": "Trinity Large Thinking",
     "provider": "openrouter",
     "family": "trinity",
-    "created_at": "2026-04-01 00:00:00 +0530",
+    "created_at": "2026-04-01 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 80000,
     "knowledge_cutoff": null,
@@ -14932,7 +32121,10 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -14943,6 +32135,42 @@
       }
     },
     "metadata": {
+      "description": "Trinity Large Thinking is a powerful open source reasoning model from the team at Arcee AI. It shows strong performance in PinchBench, agentic workloads, and reasoning tasks. Launch video: https://youtu.be/Gc82AXLa0Rg?si=4RLn6WBz33qT--B7",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 262144,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -14959,12 +32187,563 @@
       }
     }
   },
+  {
+    "id": "arcee-ai/trinity-mini",
+    "name": "Arcee AI: Trinity Mini",
+    "provider": "openrouter",
+    "family": "arcee-ai",
+    "created_at": "2025-12-01 15:08:40 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 131072,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.045,
+          "output_per_million": 0.15
+        }
+      }
+    },
+    "metadata": {
+      "description": "Trinity Mini is a 26B-parameter (3B active) sparse mixture-of-experts language model featuring 128 experts with 8 active per token. Engineered for efficient reasoning over long contexts (131k) with robust function...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 131072,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "arcee-ai/virtuoso-large",
+    "name": "Arcee AI: Virtuoso Large",
+    "provider": "openrouter",
+    "family": "arcee-ai",
+    "created_at": "2025-05-05 21:01:25 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 64000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.75,
+          "output_per_million": 1.2
+        }
+      }
+    },
+    "metadata": {
+      "description": "Virtuoso‑Large is Arcee's top‑tier general‑purpose LLM at 72 B parameters, tuned to tackle cross‑domain reasoning, creative writing and enterprise QA. Unlike many 70 B peers, it retains the 128 k...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 64000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "baidu/cobuddy:free",
+    "name": "Baidu Qianfan: CoBuddy (free)",
+    "provider": "openrouter",
+    "family": "baidu",
+    "created_at": "2026-05-06 02:44:40 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "description": "CoBuddy is a code generation model from Baidu, optimized for coding tasks and AI Agent workflows. It features high inference throughput and low end-to-end latency, with native support for tool...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "stop",
+        "tools"
+      ]
+    }
+  },
+  {
+    "id": "baidu/ernie-4.5-21b-a3b",
+    "name": "Baidu: ERNIE 4.5 21B A3B",
+    "provider": "openrouter",
+    "family": "baidu",
+    "created_at": "2025-08-12 21:29:27 UTC",
+    "context_window": 120000,
+    "max_output_tokens": 8000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.07,
+          "output_per_million": 0.28
+        }
+      }
+    },
+    "metadata": {
+      "description": "A sophisticated text-based Mixture-of-Experts (MoE) model featuring 21B total parameters with 3B activated per token, delivering exceptional multimodal understanding and generation through heterogeneous MoE structures and modality-isolated routing. Supporting an...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 120000,
+        "max_completion_tokens": 8000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "baidu/ernie-4.5-21b-a3b-thinking",
+    "name": "Baidu: ERNIE 4.5 21B A3B Thinking",
+    "provider": "openrouter",
+    "family": "baidu",
+    "created_at": "2025-10-09 22:28:07 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.07,
+          "output_per_million": 0.28
+        }
+      }
+    },
+    "metadata": {
+      "description": "ERNIE-4.5-21B-A3B-Thinking is Baidu's upgraded lightweight MoE model, refined to boost reasoning depth and quality for top-tier performance in logical puzzles, math, science, coding, text generation, and expert-level academic benchmarks.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "baidu/ernie-4.5-300b-a47b",
+    "name": "Baidu: ERNIE 4.5 300B A47B ",
+    "provider": "openrouter",
+    "family": "baidu",
+    "created_at": "2025-06-30 16:15:39 UTC",
+    "context_window": 123000,
+    "max_output_tokens": 12000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.28,
+          "output_per_million": 1.1
+        }
+      }
+    },
+    "metadata": {
+      "description": "ERNIE-4.5-300B-A47B is a 300B parameter Mixture-of-Experts (MoE) language model developed by Baidu as part of the ERNIE 4.5 series. It activates 47B parameters per token and supports text generation in...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 123000,
+        "max_completion_tokens": 12000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "baidu/ernie-4.5-vl-28b-a3b",
+    "name": "Baidu: ERNIE 4.5 VL 28B A3B",
+    "provider": "openrouter",
+    "family": "baidu",
+    "created_at": "2025-08-12 21:07:16 UTC",
+    "context_window": 30000,
+    "max_output_tokens": 8000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.14,
+          "output_per_million": 0.56
+        }
+      }
+    },
+    "metadata": {
+      "description": "A powerful multimodal Mixture-of-Experts chat model featuring 28B total parameters with 3B activated per token, delivering exceptional text and vision understanding through its innovative heterogeneous MoE structure with modality-isolated routing....",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 30000,
+        "max_completion_tokens": 8000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "baidu/ernie-4.5-vl-424b-a47b",
+    "name": "Baidu: ERNIE 4.5 VL 424B A47B ",
+    "provider": "openrouter",
+    "family": "baidu",
+    "created_at": "2025-06-30 16:28:23 UTC",
+    "context_window": 123000,
+    "max_output_tokens": 16000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.42,
+          "output_per_million": 1.25
+        }
+      }
+    },
+    "metadata": {
+      "description": "ERNIE-4.5-VL-424B-A47B is a multimodal Mixture-of-Experts (MoE) model from Baidu’s ERNIE 4.5 series, featuring 424B total parameters with 47B active per token. It is trained jointly on text and image data...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 123000,
+        "max_completion_tokens": 16000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "baidu/qianfan-ocr-fast:free",
+    "name": "Baidu: Qianfan-OCR-Fast (free)",
+    "provider": "openrouter",
+    "family": "baidu",
+    "created_at": "2026-04-20 17:51:12 UTC",
+    "context_window": 65536,
+    "max_output_tokens": 28672,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {},
+    "metadata": {
+      "description": "Qianfan-OCR-Fast is a domain-specific multimodal large model purpose-built for OCR. By leveraging specialized OCR training data while preserving versatile multimodal intelligence, it provides a powerful performance upgrade over Qianfan-OCR.",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 65536,
+        "max_completion_tokens": 28672,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "black-forest-labs/flux.2-flex",
     "name": "FLUX.2 Flex",
     "provider": "openrouter",
     "family": "flux",
-    "created_at": "2025-11-25 00:00:00 +0530",
+    "created_at": "2025-11-25 00:00:00 UTC",
     "context_window": 67344,
     "max_output_tokens": 67344,
     "knowledge_cutoff": null,
@@ -15004,7 +32783,7 @@
     "name": "FLUX.2 Klein 4B",
     "provider": "openrouter",
     "family": "flux",
-    "created_at": "2026-01-14 00:00:00 +0530",
+    "created_at": "2026-01-14 00:00:00 UTC",
     "context_window": 40960,
     "max_output_tokens": 40960,
     "knowledge_cutoff": null,
@@ -15044,7 +32823,7 @@
     "name": "FLUX.2 Max",
     "provider": "openrouter",
     "family": "flux",
-    "created_at": "2025-12-16 00:00:00 +0530",
+    "created_at": "2025-12-16 00:00:00 UTC",
     "context_window": 46864,
     "max_output_tokens": 46864,
     "knowledge_cutoff": null,
@@ -15084,7 +32863,7 @@
     "name": "FLUX.2 Pro",
     "provider": "openrouter",
     "family": "flux",
-    "created_at": "2025-11-25 00:00:00 +0530",
+    "created_at": "2025-11-25 00:00:00 UTC",
     "context_window": 46864,
     "max_output_tokens": 46864,
     "knowledge_cutoff": null,
@@ -15119,12 +32898,284 @@
       "knowledge": "2025-06"
     }
   },
+  {
+    "id": "bytedance-seed/seed-1.6",
+    "name": "ByteDance Seed: Seed 1.6",
+    "provider": "openrouter",
+    "family": "bytedance-seed",
+    "created_at": "2025-12-23 15:49:57 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.25,
+          "output_per_million": 2.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "Seed 1.6 is a general-purpose model released by the ByteDance Seed team. It incorporates multimodal capabilities and adaptive deep thinking with a 256K context window.",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "bytedance-seed/seed-1.6-flash",
+    "name": "ByteDance Seed: Seed 1.6 Flash",
+    "provider": "openrouter",
+    "family": "bytedance-seed",
+    "created_at": "2025-12-23 15:50:11 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "description": "Seed 1.6 Flash is an ultra-fast multimodal deep thinking model by ByteDance Seed, supporting both text and visual understanding. It features a 256k context window and can generate outputs of...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "bytedance-seed/seed-2.0-lite",
+    "name": "ByteDance Seed: Seed-2.0-Lite",
+    "provider": "openrouter",
+    "family": "bytedance-seed",
+    "created_at": "2026-03-10 15:40:31 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 131072,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.25,
+          "output_per_million": 2.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "Seed-2.0-Lite is a versatile, cost‑efficient enterprise workhorse that delivers strong multimodal and agent capabilities while offering noticeably lower latency, making it a practical default choice for most production workloads across...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 131072,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "bytedance-seed/seed-2.0-mini",
+    "name": "ByteDance Seed: Seed-2.0-Mini",
+    "provider": "openrouter",
+    "family": "bytedance-seed",
+    "created_at": "2026-02-26 18:38:27 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 131072,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.09999999999999999,
+          "output_per_million": 0.39999999999999997
+        }
+      }
+    },
+    "metadata": {
+      "description": "Seed-2.0-mini targets latency-sensitive, high-concurrency, and cost-sensitive scenarios, emphasizing fast response and flexible inference deployment. It delivers performance comparable to ByteDance-Seed-1.6, supports 256k context, four reasoning effort modes (minimal/low/medium/high), multimodal understanding,...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 131072,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "bytedance-seed/seedream-4.5",
     "name": "Seedream 4.5",
     "provider": "openrouter",
     "family": "seed",
-    "created_at": "2025-12-23 00:00:00 +0530",
+    "created_at": "2025-12-23 00:00:00 UTC",
     "context_window": 4096,
     "max_output_tokens": 4096,
     "knowledge_cutoff": null,
@@ -15159,12 +33210,77 @@
       "knowledge": "2025-06"
     }
   },
+  {
+    "id": "bytedance/ui-tars-1.5-7b",
+    "name": "ByteDance: UI-TARS 7B ",
+    "provider": "openrouter",
+    "family": "bytedance",
+    "created_at": "2025-07-22 17:24:16 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 2048,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.09999999999999999,
+          "output_per_million": 0.19999999999999998,
+          "cache_read_input_per_million": 0.09999999999999999
+        }
+      }
+    },
+    "metadata": {
+      "description": "UI-TARS-1.5 is a multimodal vision-language agent optimized for GUI-based environments, including desktop interfaces, web browsers, mobile systems, and games. Built by ByteDance, it builds upon the UI-TARS framework with reinforcement...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 2048,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "cognitivecomputations/dolphin-mistral-24b-venice-edition:free",
     "name": "Uncensored (free)",
     "provider": "openrouter",
     "family": "mistral",
-    "created_at": "2025-07-09 00:00:00 +0530",
+    "created_at": "2025-07-09 00:00:00 UTC",
     "context_window": 32768,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -15177,10 +33293,40 @@
       ]
     },
     "capabilities": [
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
     "pricing": {},
     "metadata": {
+      "description": "Venice Uncensored Dolphin Mistral 24B Venice Edition is a fine-tuned variant of Mistral-Small-24B-Instruct-2501, developed by dphn.ai in collaboration with Venice.ai. This model is designed as an “uncensored” instruct-tuned LLM, preserving...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -15198,12 +33344,402 @@
       "knowledge": "2025-06"
     }
   },
+  {
+    "id": "cohere/command-a",
+    "name": "Cohere: Command A",
+    "provider": "openrouter",
+    "family": "cohere",
+    "created_at": "2025-03-13 19:32:22 UTC",
+    "context_window": 256000,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "Command A is an open-weights 111B parameter model with a 256k context window focused on delivering great performance across agentic, multilingual, and coding use cases. Compared to other leading proprietary...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 256000,
+        "max_completion_tokens": 8192,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "cohere/command-r-08-2024",
+    "name": "Cohere: Command R (08-2024)",
+    "provider": "openrouter",
+    "family": "cohere",
+    "created_at": "2024-08-30 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.15,
+          "output_per_million": 0.6
+        }
+      }
+    },
+    "metadata": {
+      "description": "command-r-08-2024 is an update of the [Command R](/models/cohere/command-r) with improved performance for multilingual retrieval-augmented generation (RAG) and tool use. More broadly, it is better at math, code and reasoning and...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Cohere",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 4000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "cohere/command-r-plus-08-2024",
+    "name": "Cohere: Command R+ (08-2024)",
+    "provider": "openrouter",
+    "family": "cohere",
+    "created_at": "2024-08-30 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "command-r-plus-08-2024 is an update of the [Command R+](/models/cohere/command-r-plus) with roughly 50% higher throughput and 25% lower latencies as compared to the previous Command R+ version, while keeping the hardware footprint...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Cohere",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 4000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "cohere/command-r7b-12-2024",
+    "name": "Cohere: Command R7B (12-2024)",
+    "provider": "openrouter",
+    "family": "cohere",
+    "created_at": "2024-12-14 06:35:52 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.0375,
+          "output_per_million": 0.15
+        }
+      }
+    },
+    "metadata": {
+      "description": "Command R7B (12-2024) is a small, fast update of the Command R+ model, delivered in December 2024. It excels at RAG, tool use, agents, and similar tasks requiring complex reasoning...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Cohere",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 4000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "deepcogito/cogito-v2.1-671b",
+    "name": "Deep Cogito: Cogito v2.1 671B",
+    "provider": "openrouter",
+    "family": "deepcogito",
+    "created_at": "2025-11-13 22:00:33 UTC",
+    "context_window": 128000,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 1.25
+        }
+      }
+    },
+    "metadata": {
+      "description": "Cogito v2.1 671B MoE represents one of the strongest open models globally, matching performance of frontier closed and open models. This model is trained using self play with reinforcement learning...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "deepseek/deepseek-chat",
+    "name": "DeepSeek: DeepSeek V3",
+    "provider": "openrouter",
+    "family": "deepseek",
+    "created_at": "2024-12-26 19:28:40 UTC",
+    "context_window": 163840,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.32,
+          "output_per_million": 0.8899999999999999
+        }
+      }
+    },
+    "metadata": {
+      "description": "DeepSeek-V3 is the latest model from the DeepSeek team, building upon the instruction following and coding abilities of the previous versions. Pre-trained on nearly 15 trillion tokens, the reported evaluations...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "DeepSeek",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 163840,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "deepseek/deepseek-chat-v3-0324",
     "name": "DeepSeek V3 0324",
     "provider": "openrouter",
     "family": "deepseek",
-    "created_at": "2025-03-24 00:00:00 +0530",
+    "created_at": "2025-03-24 00:00:00 UTC",
     "context_window": 16384,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -15216,10 +33752,56 @@
       ]
     },
     "capabilities": [
-      "structured_output"
+      "structured_output",
+      "streaming",
+      "function_calling",
+      "predicted_outputs"
     ],
-    "pricing": {},
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.19999999999999998,
+          "output_per_million": 0.77,
+          "cache_read_input_per_million": 0.135
+        }
+      }
+    },
     "metadata": {
+      "description": "DeepSeek V3, a 685B-parameter, mixture-of-experts model, is the latest iteration of the flagship chat model family from the DeepSeek team. It succeeds the [DeepSeek V3](/deepseek/deepseek-chat-v3) model and performs really well...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "DeepSeek",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 163840,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -15242,7 +33824,7 @@
     "name": "DeepSeek-V3.1",
     "provider": "openrouter",
     "family": "deepseek",
-    "created_at": "2025-08-21 00:00:00 +0530",
+    "created_at": "2025-08-21 00:00:00 UTC",
     "context_window": 163840,
     "max_output_tokens": 163840,
     "knowledge_cutoff": null,
@@ -15257,7 +33839,9 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -15268,6 +33852,45 @@
       }
     },
     "metadata": {
+      "description": "DeepSeek-V3.1 is a large hybrid reasoning model (671B parameters, 37B active) that supports both thinking and non-thinking modes via prompt templates. It extends the DeepSeek-V3 base with a two-phase long-context...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "DeepSeek",
+        "instruct_type": "deepseek-v3.1"
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": 7168,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -15290,7 +33913,7 @@
     "name": "DeepSeek: R1",
     "provider": "openrouter",
     "family": "deepseek-thinking",
-    "created_at": "2025-01-20 00:00:00 +0530",
+    "created_at": "2025-01-20 00:00:00 UTC",
     "context_window": 64000,
     "max_output_tokens": 16000,
     "knowledge_cutoff": null,
@@ -15304,7 +33927,8 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -15315,6 +33939,40 @@
       }
     },
     "metadata": {
+      "description": "DeepSeek R1 is here: Performance on par with [OpenAI o1](/openai/o1), but open-sourced and with fully open reasoning tokens. It's 671B parameters in size, with 37B active in an inference pass....",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "DeepSeek",
+        "instruct_type": "deepseek-r1"
+      },
+      "top_provider": {
+        "context_length": 64000,
+        "max_completion_tokens": 16000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -15332,12 +33990,84 @@
       "knowledge": "2024-07"
     }
   },
+  {
+    "id": "deepseek/deepseek-r1-0528",
+    "name": "DeepSeek: R1 0528",
+    "provider": "openrouter",
+    "family": "deepseek",
+    "created_at": "2025-05-28 17:59:30 UTC",
+    "context_window": 163840,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 2.1500000000000004,
+          "cache_read_input_per_million": 0.35
+        }
+      }
+    },
+    "metadata": {
+      "description": "May 28th update to the [original DeepSeek R1](/deepseek/deepseek-r1) Performance on par with [OpenAI o1](/openai/o1), but open-sourced and with fully open reasoning tokens. It's 671B parameters in size, with 37B active...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "DeepSeek",
+        "instruct_type": "deepseek-r1"
+      },
+      "top_provider": {
+        "context_length": 163840,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "deepseek/deepseek-r1-distill-llama-70b",
     "name": "DeepSeek R1 Distill Llama 70B",
     "provider": "openrouter",
     "family": "deepseek-thinking",
-    "created_at": "2025-01-23 00:00:00 +0530",
+    "created_at": "2025-01-23 00:00:00 UTC",
     "context_window": 8192,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -15351,10 +34081,53 @@
     },
     "capabilities": [
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
-    "pricing": {},
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.7,
+          "output_per_million": 0.7999999999999999
+        }
+      }
+    },
     "metadata": {
+      "description": "DeepSeek R1 Distill Llama 70B is a distilled large language model based on [Llama-3.3-70B-Instruct](/meta-llama/llama-3.3-70b-instruct), using outputs from [DeepSeek R1](/deepseek/deepseek-r1). The model combines advanced distillation techniques to achieve high performance across...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "deepseek-r1"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -15372,12 +34145,78 @@
       "knowledge": "2024-10"
     }
   },
+  {
+    "id": "deepseek/deepseek-r1-distill-qwen-32b",
+    "name": "DeepSeek: R1 Distill Qwen 32B",
+    "provider": "openrouter",
+    "family": "deepseek",
+    "created_at": "2025-01-29 23:53:50 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.29,
+          "output_per_million": 0.29
+        }
+      }
+    },
+    "metadata": {
+      "description": "DeepSeek R1 Distill Qwen 32B is a distilled large language model based on [Qwen 2.5 32B](https://huggingface.co/Qwen/Qwen2.5-32B), using outputs from [DeepSeek R1](/deepseek/deepseek-r1). It outperforms OpenAI's o1-mini across various benchmarks, achieving new...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen",
+        "instruct_type": "deepseek-r1"
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "deepseek/deepseek-v3.1-terminus",
     "name": "DeepSeek V3.1 Terminus",
     "provider": "openrouter",
     "family": "deepseek",
-    "created_at": "2025-09-22 00:00:00 +0530",
+    "created_at": "2025-09-22 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -15392,7 +34231,9 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -15403,6 +34244,43 @@
       }
     },
     "metadata": {
+      "description": "DeepSeek-V3.1 Terminus is an update to [DeepSeek V3.1](/deepseek/deepseek-chat-v3.1) that maintains the model's original capabilities while addressing issues reported by users, including language consistency and agent capabilities, further optimizing the model's...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "DeepSeek",
+        "instruct_type": "deepseek-v3.1"
+      },
+      "top_provider": {
+        "context_length": 163840,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -15425,7 +34303,7 @@
     "name": "DeepSeek V3.1 Terminus (exacto)",
     "provider": "openrouter",
     "family": "deepseek",
-    "created_at": "2025-09-22 00:00:00 +0530",
+    "created_at": "2025-09-22 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -15473,7 +34351,7 @@
     "name": "DeepSeek V3.2",
     "provider": "openrouter",
     "family": "deepseek",
-    "created_at": "2025-12-01 00:00:00 +0530",
+    "created_at": "2025-12-01 00:00:00 UTC",
     "context_window": 163840,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -15488,7 +34366,9 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -15499,6 +34379,43 @@
       }
     },
     "metadata": {
+      "description": "DeepSeek-V3.2 is a large language model designed to harmonize high computational efficiency with strong reasoning and agentic tool-use performance. It introduces DeepSeek Sparse Attention (DSA), a fine-grained sparse attention mechanism...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "DeepSeek",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -15516,12 +34433,83 @@
       "knowledge": "2024-07"
     }
   },
+  {
+    "id": "deepseek/deepseek-v3.2-exp",
+    "name": "DeepSeek: DeepSeek V3.2 Exp",
+    "provider": "openrouter",
+    "family": "deepseek",
+    "created_at": "2025-09-29 12:54:41 UTC",
+    "context_window": 163840,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.27,
+          "output_per_million": 0.41
+        }
+      }
+    },
+    "metadata": {
+      "description": "DeepSeek-V3.2-Exp is an experimental large language model released by DeepSeek as an intermediate step between V3.1 and future architectures. It introduces DeepSeek Sparse Attention (DSA), a fine-grained sparse attention mechanism...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "DeepSeek",
+        "instruct_type": "deepseek-v3.1"
+      },
+      "top_provider": {
+        "context_length": 163840,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "deepseek/deepseek-v3.2-speciale",
     "name": "DeepSeek V3.2 Speciale",
     "provider": "openrouter",
     "family": "deepseek",
-    "created_at": "2025-12-01 00:00:00 +0530",
+    "created_at": "2025-12-01 00:00:00 UTC",
     "context_window": 163840,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -15536,7 +34524,9 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -15547,6 +34537,41 @@
       }
     },
     "metadata": {
+      "description": "DeepSeek-V3.2-Speciale is a high-compute variant of DeepSeek-V3.2 optimized for maximum reasoning and agentic performance. It builds on DeepSeek Sparse Attention (DSA) for efficient long-context processing, then scales post-training reinforcement learning...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "DeepSeek",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 163840,
+        "max_completion_tokens": 163840,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -15564,12 +34589,268 @@
       "knowledge": "2024-07"
     }
   },
+  {
+    "id": "deepseek/deepseek-v4-flash",
+    "name": "DeepSeek V4 Flash",
+    "provider": "openrouter",
+    "family": "deepseek-flash",
+    "created_at": "2026-04-24 00:00:00 UTC",
+    "context_window": 1048576,
+    "max_output_tokens": 393216,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.14,
+          "output_per_million": 0.28,
+          "cache_read_input_per_million": 0.028
+        }
+      }
+    },
+    "metadata": {
+      "description": "DeepSeek V4 Flash is an efficiency-optimized Mixture-of-Experts model from DeepSeek with 284B total parameters and 13B activated parameters, supporting a 1M-token context window. It is designed for fast inference and...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "DeepSeek",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 384000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ],
+      "source": "models.dev",
+      "provider_id": "openrouter",
+      "open_weights": true,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2026-04-24",
+      "interleaved": {
+        "field": "reasoning_content"
+      },
+      "cost": {
+        "input": 0.14,
+        "output": 0.28,
+        "cache_read": 0.028
+      },
+      "limit": {
+        "context": 1048576,
+        "output": 393216
+      },
+      "knowledge": "2025-05"
+    }
+  },
+  {
+    "id": "deepseek/deepseek-v4-pro",
+    "name": "DeepSeek V4 Pro",
+    "provider": "openrouter",
+    "family": "deepseek-thinking",
+    "created_at": "2026-04-24 00:00:00 UTC",
+    "context_window": 1048576,
+    "max_output_tokens": 393216,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.74,
+          "output_per_million": 3.48,
+          "cache_read_input_per_million": 0.145
+        }
+      }
+    },
+    "metadata": {
+      "description": "DeepSeek V4 Pro is a large-scale Mixture-of-Experts model from DeepSeek with 1.6T total parameters and 49B activated parameters, supporting a 1M-token context window. It is designed for advanced reasoning, coding,...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "DeepSeek",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 384000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ],
+      "source": "models.dev",
+      "provider_id": "openrouter",
+      "open_weights": true,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2026-04-24",
+      "interleaved": {
+        "field": "reasoning_content"
+      },
+      "cost": {
+        "input": 1.74,
+        "output": 3.48,
+        "cache_read": 0.145
+      },
+      "limit": {
+        "context": 1048576,
+        "output": 393216
+      },
+      "knowledge": "2025-05"
+    }
+  },
+  {
+    "id": "essentialai/rnj-1-instruct",
+    "name": "EssentialAI: Rnj 1 Instruct",
+    "provider": "openrouter",
+    "family": "essentialai",
+    "created_at": "2025-12-07 08:07:27 UTC",
+    "context_window": 32768,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.15,
+          "output_per_million": 0.15
+        }
+      }
+    },
+    "metadata": {
+      "description": "Rnj-1 is an 8B-parameter, dense, open-weight model family developed by Essential AI and trained from scratch with a focus on programming, math, and scientific reasoning. The model demonstrates strong performance...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "google/gemini-2.0-flash-001",
     "name": "Gemini 2.0 Flash",
     "provider": "openrouter",
     "family": "gemini-flash",
-    "created_at": "2024-12-11 00:00:00 +0530",
+    "created_at": "2024-12-11 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -15588,18 +34869,52 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.1,
           "output_per_million": 0.4,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       }
     },
     "metadata": {
+      "description": "Gemini Flash 2.0 offers a significantly faster time to first token (TTFT) compared to [Gemini Flash 1.5](/google/gemini-flash-1.5), while maintaining quality on par with larger models like [Gemini Pro 1.5](/google/gemini-pro-1.5). It...",
+      "architecture": {
+        "modality": "text+image+file+audio+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file",
+          "audio",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 8192,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -15618,12 +34933,83 @@
       "knowledge": "2024-06"
     }
   },
+  {
+    "id": "google/gemini-2.0-flash-lite-001",
+    "name": "Google: Gemini 2.0 Flash Lite",
+    "provider": "openrouter",
+    "family": "google",
+    "created_at": "2025-02-25 17:56:52 UTC",
+    "context_window": 1048576,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "file",
+        "audio",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3,
+          "reasoning_output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "description": "Gemini 2.0 Flash Lite offers a significantly faster time to first token (TTFT) compared to [Gemini Flash 1.5](/google/gemini-flash-1.5), while maintaining quality on par with larger models like [Gemini Pro 1.5](/google/gemini-pro-1.5),...",
+      "architecture": {
+        "modality": "text+image+file+audio+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file",
+          "audio",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 8192,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "google/gemini-2.5-flash",
     "name": "Gemini 2.5 Flash",
     "provider": "openrouter",
     "family": "gemini-flash",
-    "created_at": "2025-07-17 00:00:00 +0530",
+    "created_at": "2025-07-17 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -15643,18 +35029,54 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.3,
           "output_per_million": 2.5,
-          "cached_input_per_million": 0.0375
+          "cache_read_input_per_million": 0.0375
         }
       }
     },
     "metadata": {
+      "description": "Gemini 2.5 Flash is Google's state-of-the-art workhorse model, specifically designed for advanced reasoning, coding, mathematics, and scientific tasks. It includes built-in \"thinking\" capabilities, enabling it to provide responses with greater...",
+      "architecture": {
+        "modality": "text+image+file+audio+video->text",
+        "input_modalities": [
+          "file",
+          "image",
+          "text",
+          "audio",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 65535,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -15673,12 +35095,77 @@
       "knowledge": "2025-01"
     }
   },
+  {
+    "id": "google/gemini-2.5-flash-image",
+    "name": "Google: Nano Banana (Gemini 2.5 Flash Image)",
+    "provider": "openrouter",
+    "family": "google",
+    "created_at": "2025-10-07 20:53:51 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text"
+      ],
+      "output": [
+        "image",
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.3,
+          "output_per_million": 2.5,
+          "cache_read_input_per_million": 0.03,
+          "reasoning_output_per_million": 2.5
+        }
+      }
+    },
+    "metadata": {
+      "description": "Gemini 2.5 Flash Image, a.k.a. \"Nano Banana,\" is now generally available. It is a state of the art image generation model with contextual understanding. It is capable of image generation,...",
+      "architecture": {
+        "modality": "text+image->text+image",
+        "input_modalities": [
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "image",
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "google/gemini-2.5-flash-lite",
     "name": "Gemini 2.5 Flash Lite",
     "provider": "openrouter",
     "family": "gemini-flash-lite",
-    "created_at": "2025-06-17 00:00:00 +0530",
+    "created_at": "2025-06-17 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -15698,18 +35185,54 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.1,
           "output_per_million": 0.4,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       }
     },
     "metadata": {
+      "description": "Gemini 2.5 Flash-Lite is a lightweight reasoning model in the Gemini 2.5 family, optimized for ultra-low latency and cost efficiency. It offers improved throughput, faster token generation, and better performance...",
+      "architecture": {
+        "modality": "text+image+file+audio+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file",
+          "audio",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 65535,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -15733,7 +35256,7 @@
     "name": "Gemini 2.5 Flash Lite Preview 09-25",
     "provider": "openrouter",
     "family": "gemini-flash-lite",
-    "created_at": "2025-09-25 00:00:00 +0530",
+    "created_at": "2025-09-25 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -15753,18 +35276,54 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.1,
           "output_per_million": 0.4,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       }
     },
     "metadata": {
+      "description": "Gemini 2.5 Flash-Lite is a lightweight reasoning model in the Gemini 2.5 family, optimized for ultra-low latency and cost efficiency. It offers improved throughput, faster token generation, and better performance...",
+      "architecture": {
+        "modality": "text+image+file+audio+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file",
+          "audio",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 65535,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -15788,7 +35347,7 @@
     "name": "Gemini 2.5 Flash Preview 09-25",
     "provider": "openrouter",
     "family": "gemini-flash",
-    "created_at": "2025-09-25 00:00:00 +0530",
+    "created_at": "2025-09-25 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -15815,7 +35374,7 @@
         "standard": {
           "input_per_million": 0.3,
           "output_per_million": 2.5,
-          "cached_input_per_million": 0.031
+          "cache_read_input_per_million": 0.031
         }
       }
     },
@@ -15843,7 +35402,7 @@
     "name": "Gemini 2.5 Pro",
     "provider": "openrouter",
     "family": "gemini-pro",
-    "created_at": "2025-03-20 00:00:00 +0530",
+    "created_at": "2025-03-20 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -15863,18 +35422,54 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.31
+          "cache_read_input_per_million": 0.125
         }
       }
     },
     "metadata": {
+      "description": "Gemini 2.5 Pro is Google’s state-of-the-art AI model designed for advanced reasoning, coding, mathematics, and scientific tasks. It employs “thinking” capabilities, enabling it to reason through responses with enhanced accuracy...",
+      "architecture": {
+        "modality": "text+image+file+audio+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file",
+          "audio",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -15884,7 +35479,12 @@
       "cost": {
         "input": 1.25,
         "output": 10,
-        "cache_read": 0.31
+        "cache_read": 0.125,
+        "context_over_200k": {
+          "input": 2.5,
+          "output": 15,
+          "cache_read": 0.25
+        }
       },
       "limit": {
         "context": 1048576,
@@ -15893,12 +35493,84 @@
       "knowledge": "2025-01"
     }
   },
+  {
+    "id": "google/gemini-2.5-pro-preview",
+    "name": "Google: Gemini 2.5 Pro Preview 06-05",
+    "provider": "openrouter",
+    "family": "google",
+    "created_at": "2025-06-05 15:27:37 UTC",
+    "context_window": 1048576,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "file",
+        "image",
+        "text",
+        "audio"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 0.125,
+          "reasoning_output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "Gemini 2.5 Pro is Google’s state-of-the-art AI model designed for advanced reasoning, coding, mathematics, and scientific tasks. It employs “thinking” capabilities, enabling it to reason through responses with enhanced accuracy...",
+      "architecture": {
+        "modality": "text+image+file+audio->text",
+        "input_modalities": [
+          "file",
+          "image",
+          "text",
+          "audio"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "google/gemini-2.5-pro-preview-05-06",
     "name": "Gemini 2.5 Pro Preview 05-06",
     "provider": "openrouter",
     "family": "gemini-pro",
-    "created_at": "2025-05-06 00:00:00 +0530",
+    "created_at": "2025-05-06 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -15918,18 +35590,54 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.31
+          "cache_read_input_per_million": 0.31
         }
       }
     },
     "metadata": {
+      "description": "Gemini 2.5 Pro is Google’s state-of-the-art AI model designed for advanced reasoning, coding, mathematics, and scientific tasks. It employs “thinking” capabilities, enabling it to reason through responses with enhanced accuracy...",
+      "architecture": {
+        "modality": "text+image+file+audio+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file",
+          "audio",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 65535,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -15953,7 +35661,7 @@
     "name": "Gemini 2.5 Pro Preview 06-05",
     "provider": "openrouter",
     "family": "gemini-pro",
-    "created_at": "2025-06-05 00:00:00 +0530",
+    "created_at": "2025-06-05 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -15980,7 +35688,7 @@
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.31
+          "cache_read_input_per_million": 0.31
         }
       }
     },
@@ -16008,7 +35716,7 @@
     "name": "Gemini 3 Flash Preview",
     "provider": "openrouter",
     "family": "gemini-flash",
-    "created_at": "2025-12-17 00:00:00 +0530",
+    "created_at": "2025-12-17 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -16028,18 +35736,54 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.5,
           "output_per_million": 3,
-          "cached_input_per_million": 0.05
+          "cache_read_input_per_million": 0.05
         }
       }
     },
     "metadata": {
+      "description": "Gemini 3 Flash Preview is a high speed, high value thinking model designed for agentic workflows, multi turn chat, and coding assistance. It delivers near Pro level reasoning and tool...",
+      "architecture": {
+        "modality": "text+image+file+audio+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file",
+          "audio",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -16061,12 +35805,79 @@
       "knowledge": "2025-01"
     }
   },
+  {
+    "id": "google/gemini-3-pro-image-preview",
+    "name": "Google: Nano Banana Pro (Gemini 3 Pro Image Preview)",
+    "provider": "openrouter",
+    "family": "google",
+    "created_at": "2025-11-20 15:49:57 UTC",
+    "context_window": 65536,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text"
+      ],
+      "output": [
+        "image",
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "output_per_million": 12.0,
+          "cache_read_input_per_million": 0.19999999999999998,
+          "reasoning_output_per_million": 12.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "Nano Banana Pro is Google’s most advanced image-generation and editing model, built on Gemini 3 Pro. It extends the original Nano Banana with significantly improved multimodal reasoning, real-world grounding, and...",
+      "architecture": {
+        "modality": "text+image->text+image",
+        "input_modalities": [
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "image",
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 65536,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "google/gemini-3-pro-preview",
     "name": "Gemini 3 Pro Preview",
     "provider": "openrouter",
     "family": "gemini-pro",
-    "created_at": "2025-11-18 00:00:00 +0530",
+    "created_at": "2025-11-18 00:00:00 UTC",
     "context_window": 1050000,
     "max_output_tokens": 66000,
     "knowledge_cutoff": null,
@@ -16117,12 +35928,94 @@
       "knowledge": "2025-01"
     }
   },
+  {
+    "id": "google/gemini-3.1-flash-image-preview",
+    "name": "Gemini 3.1 Flash Image Preview (Nano Banana 2)",
+    "provider": "openrouter",
+    "family": "gemini-flash",
+    "created_at": "2026-02-26 00:00:00 UTC",
+    "context_window": 65536,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text",
+        "image"
+      ]
+    },
+    "capabilities": [
+      "structured_output",
+      "reasoning",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 3
+        }
+      }
+    },
+    "metadata": {
+      "description": "Gemini 3.1 Flash Image Preview, a.k.a. \"Nano Banana 2,\" is Google’s latest state of the art image generation and editing model, delivering Pro-level visual quality at Flash speed. It combines...",
+      "architecture": {
+        "modality": "text+image->text+image",
+        "input_modalities": [
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "image",
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 65536,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_p"
+      ],
+      "source": "models.dev",
+      "provider_id": "openrouter",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2026-02-26",
+      "cost": {
+        "input": 0.5,
+        "output": 3
+      },
+      "limit": {
+        "context": 65536,
+        "output": 65536
+      },
+      "knowledge": "2025-01"
+    }
+  },
   {
     "id": "google/gemini-3.1-flash-lite-preview",
     "name": "Gemini 3.1 Flash Lite Preview",
     "provider": "openrouter",
     "family": "gemini-flash-lite",
-    "created_at": "2026-03-03 00:00:00 +0530",
+    "created_at": "2026-03-03 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -16142,14 +36035,16 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.25,
           "output_per_million": 1.5,
-          "cached_input_per_million": 0.025,
+          "cache_read_input_per_million": 0.025,
+          "cache_write_input_per_million": 0.083,
           "reasoning_output_per_million": 1.5
         }
       },
@@ -16161,6 +36056,41 @@
       }
     },
     "metadata": {
+      "description": "Gemini 3.1 Flash Lite Preview is Google's high-efficiency model optimized for high-volume use cases. It outperforms Gemini 2.5 Flash Lite on overall quality and approaches Gemini 2.5 Flash performance across...",
+      "architecture": {
+        "modality": "text+image+file+audio+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "video",
+          "file",
+          "audio"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -16187,7 +36117,7 @@
     "name": "Gemini 3.1 Pro Preview",
     "provider": "openrouter",
     "family": "gemini-pro",
-    "created_at": "2026-02-19 00:00:00 +0530",
+    "created_at": "2026-02-19 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -16207,7 +36137,8 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -16219,6 +36150,41 @@
       }
     },
     "metadata": {
+      "description": "Gemini 3.1 Pro Preview is Google’s frontier reasoning model, delivering enhanced software engineering performance, improved agentic reliability, and more efficient token usage across complex workflows. Building on the multimodal foundation...",
+      "architecture": {
+        "modality": "text+image+file+audio+video->text",
+        "input_modalities": [
+          "audio",
+          "file",
+          "image",
+          "text",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -16250,7 +36216,7 @@
     "name": "Gemini 3.1 Pro Preview Custom Tools",
     "provider": "openrouter",
     "family": "gemini-pro",
-    "created_at": "2026-02-19 00:00:00 +0530",
+    "created_at": "2026-02-19 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -16270,7 +36236,8 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -16282,6 +36249,41 @@
       }
     },
     "metadata": {
+      "description": "Gemini 3.1 Pro Preview Custom Tools is a variant of Gemini 3.1 Pro that improves tool selection behavior by preventing overuse of a general bash tool when more efficient third-party...",
+      "architecture": {
+        "modality": "text+image+file+audio+video->text",
+        "input_modalities": [
+          "text",
+          "audio",
+          "image",
+          "video",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -16308,12 +36310,74 @@
       "knowledge": "2025-01"
     }
   },
+  {
+    "id": "google/gemma-2-27b-it",
+    "name": "Google: Gemma 2 27B",
+    "provider": "openrouter",
+    "family": "google",
+    "created_at": "2024-07-13 00:00:00 UTC",
+    "context_window": 8192,
+    "max_output_tokens": 2048,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.65,
+          "output_per_million": 0.65
+        }
+      }
+    },
+    "metadata": {
+      "description": "Gemma 2 27B by Google is an open model built from the same research and technology used to create the [Gemini models](/models?q=gemini). Gemma models are well-suited for a variety of...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": "gemma"
+      },
+      "top_provider": {
+        "context_length": 8192,
+        "max_completion_tokens": 2048,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "google/gemma-2-9b-it",
     "name": "Gemma 2 9B",
     "provider": "openrouter",
     "family": "gemma",
-    "created_at": "2024-06-28 00:00:00 +0530",
+    "created_at": "2024-06-28 00:00:00 UTC",
     "context_window": 8192,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -16357,7 +36421,7 @@
     "name": "Gemma 3 12B",
     "provider": "openrouter",
     "family": "gemma",
-    "created_at": "2025-03-13 00:00:00 +0530",
+    "created_at": "2025-03-13 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -16372,7 +36436,10 @@
     },
     "capabilities": [
       "structured_output",
-      "vision"
+      "vision",
+      "streaming",
+      "function_calling",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -16383,6 +36450,42 @@
       }
     },
     "metadata": {
+      "description": "Gemma 3 introduces multimodality, supporting vision-language input and text outputs. It handles context windows up to 128k tokens, understands over 140 languages, and offers improved math, reasoning, and chat capabilities,...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": "gemma"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -16405,7 +36508,7 @@
     "name": "Gemma 3 12B (free)",
     "provider": "openrouter",
     "family": "gemma",
-    "created_at": "2025-03-13 00:00:00 +0530",
+    "created_at": "2025-03-13 00:00:00 UTC",
     "context_window": 32768,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -16445,7 +36548,7 @@
     "name": "Gemma 3 27B",
     "provider": "openrouter",
     "family": "gemma",
-    "created_at": "2025-03-12 00:00:00 +0530",
+    "created_at": "2025-03-12 00:00:00 UTC",
     "context_window": 96000,
     "max_output_tokens": 96000,
     "knowledge_cutoff": null,
@@ -16461,7 +36564,9 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision"
+      "vision",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -16472,6 +36577,42 @@
       }
     },
     "metadata": {
+      "description": "Gemma 3 introduces multimodality, supporting vision-language input and text outputs. It handles context windows up to 128k tokens, understands over 140 languages, and offers improved math, reasoning, and chat capabilities,...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": "gemma"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -16494,7 +36635,7 @@
     "name": "Gemma 3 27B (free)",
     "provider": "openrouter",
     "family": "gemma",
-    "created_at": "2025-03-12 00:00:00 +0530",
+    "created_at": "2025-03-12 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -16535,7 +36676,7 @@
     "name": "Gemma 3 4B",
     "provider": "openrouter",
     "family": "gemma",
-    "created_at": "2025-03-13 00:00:00 +0530",
+    "created_at": "2025-03-13 00:00:00 UTC",
     "context_window": 96000,
     "max_output_tokens": 96000,
     "knowledge_cutoff": null,
@@ -16549,7 +36690,10 @@
       ]
     },
     "capabilities": [
-      "vision"
+      "vision",
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -16560,6 +36704,40 @@
       }
     },
     "metadata": {
+      "description": "Gemma 3 introduces multimodality, supporting vision-language input and text outputs. It handles context windows up to 128k tokens, understands over 140 languages, and offers improved math, reasoning, and chat capabilities,...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemini",
+        "instruct_type": "gemma"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -16582,7 +36760,7 @@
     "name": "Gemma 3 4B (free)",
     "provider": "openrouter",
     "family": "gemma",
-    "created_at": "2025-03-13 00:00:00 +0530",
+    "created_at": "2025-03-13 00:00:00 UTC",
     "context_window": 32768,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -16622,7 +36800,7 @@
     "name": "Gemma 3n 2B (free)",
     "provider": "openrouter",
     "family": "gemma",
-    "created_at": "2025-07-09 00:00:00 +0530",
+    "created_at": "2025-07-09 00:00:00 UTC",
     "context_window": 8192,
     "max_output_tokens": 2000,
     "knowledge_cutoff": null,
@@ -16659,7 +36837,7 @@
     "name": "Gemma 3n 4B",
     "provider": "openrouter",
     "family": "gemma",
-    "created_at": "2025-05-20 00:00:00 +0530",
+    "created_at": "2025-05-20 00:00:00 UTC",
     "context_window": 32768,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -16671,7 +36849,10 @@
         "text"
       ]
     },
-    "capabilities": [],
+    "capabilities": [
+      "streaming",
+      "predicted_outputs"
+    ],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -16681,6 +36862,36 @@
       }
     },
     "metadata": {
+      "description": "Gemma 3n E4B-it is optimized for efficient execution on mobile and low-resource devices, such as phones, laptops, and tablets. It supports multimodal inputs—including text, visual data, and audio—enabling diverse tasks...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -16703,7 +36914,7 @@
     "name": "Gemma 3n 4B (free)",
     "provider": "openrouter",
     "family": "gemma",
-    "created_at": "2025-05-20 00:00:00 +0530",
+    "created_at": "2025-05-20 00:00:00 UTC",
     "context_window": 8192,
     "max_output_tokens": 2000,
     "knowledge_cutoff": null,
@@ -16740,7 +36951,7 @@
     "name": "Gemma 4 26B A4B",
     "provider": "openrouter",
     "family": "gemma",
-    "created_at": "2026-04-03 00:00:00 +0530",
+    "created_at": "2026-04-03 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 262144,
     "knowledge_cutoff": null,
@@ -16758,7 +36969,9 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -16769,6 +36982,47 @@
       }
     },
     "metadata": {
+      "description": "Gemma 4 26B A4B IT is an instruction-tuned Mixture-of-Experts (MoE) model from Google DeepMind. Despite 25.2B total parameters, only 3.8B activate per token during inference — delivering near-31B quality at...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemma",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -16791,7 +37045,7 @@
     "name": "Gemma 4 26B A4B (free)",
     "provider": "openrouter",
     "family": "gemma",
-    "created_at": "2026-04-03 00:00:00 +0530",
+    "created_at": "2026-04-03 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -16809,10 +37063,42 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {},
     "metadata": {
+      "description": "Gemma 4 26B A4B IT is an instruction-tuned Mixture-of-Experts (MoE) model from Google DeepMind. Despite 25.2B total parameters, only 3.8B activate per token during inference — delivering near-31B quality at...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemma",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -16835,7 +37121,7 @@
     "name": "Gemma 4 31B",
     "provider": "openrouter",
     "family": "gemma",
-    "created_at": "2026-04-02 00:00:00 +0530",
+    "created_at": "2026-04-02 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 262144,
     "knowledge_cutoff": null,
@@ -16853,7 +37139,9 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -16864,6 +37152,47 @@
       }
     },
     "metadata": {
+      "description": "Gemma 4 31B Instruct is Google DeepMind's 30.7B dense multimodal model supporting text and image input with text output. Features a 256K token context window, configurable thinking/reasoning mode, native function...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemma",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -16886,7 +37215,7 @@
     "name": "Gemma 4 31B (free)",
     "provider": "openrouter",
     "family": "gemma",
-    "created_at": "2026-04-02 00:00:00 +0530",
+    "created_at": "2026-04-02 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -16904,10 +37233,42 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {},
     "metadata": {
+      "description": "Gemma 4 31B Instruct is Google DeepMind's 30.7B dense multimodal model supporting text and image input with text output. Features a 256K token context window, configurable thinking/reasoning mode, native function...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Gemma",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -16925,12 +37286,315 @@
       "knowledge": "2025-01"
     }
   },
+  {
+    "id": "google/lyria-3-clip-preview",
+    "name": "Google: Lyria 3 Clip Preview",
+    "provider": "openrouter",
+    "family": "google",
+    "created_at": "2026-03-30 21:47:35 UTC",
+    "context_window": 1048576,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text",
+        "audio"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {},
+    "metadata": {
+      "description": "30 second duration clips are priced at $0.04 per clip. Lyria 3 is Google's family of music generation models, available through the Gemini API. With Lyria 3, you can generate...",
+      "architecture": {
+        "modality": "text+image->text+audio",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text",
+          "audio"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "response_format",
+        "seed",
+        "temperature",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "google/lyria-3-pro-preview",
+    "name": "Google: Lyria 3 Pro Preview",
+    "provider": "openrouter",
+    "family": "google",
+    "created_at": "2026-03-30 21:48:06 UTC",
+    "context_window": 1048576,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text",
+        "audio"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {},
+    "metadata": {
+      "description": "Full-length songs are priced at $0.08 per song. Lyria 3 is Google's family of music generation models, available through the Gemini API. With Lyria 3, you can generate high-quality, 48kHz...",
+      "architecture": {
+        "modality": "text+image->text+audio",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text",
+          "audio"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "response_format",
+        "seed",
+        "temperature",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "gryphe/mythomax-l2-13b",
+    "name": "MythoMax 13B",
+    "provider": "openrouter",
+    "family": "gryphe",
+    "created_at": "2023-07-02 00:00:00 UTC",
+    "context_window": 4096,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.06,
+          "output_per_million": 0.06
+        }
+      }
+    },
+    "metadata": {
+      "description": "One of the highest performing and most popular fine-tunes of Llama 2 13B, with rich descriptions and roleplay. #merge",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama2",
+        "instruct_type": "alpaca"
+      },
+      "top_provider": {
+        "context_length": 4096,
+        "max_completion_tokens": 4096,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_a",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "ibm-granite/granite-4.0-h-micro",
+    "name": "IBM: Granite 4.0 Micro",
+    "provider": "openrouter",
+    "family": "ibm-granite",
+    "created_at": "2025-10-20 02:34:55 UTC",
+    "context_window": 131000,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.017,
+          "output_per_million": 0.11
+        }
+      }
+    },
+    "metadata": {
+      "description": "Granite-4.0-H-Micro is a 3B parameter from the Granite 4 family of models. These models are the latest in a series of models released by IBM. They are fine-tuned for long...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "ibm-granite/granite-4.1-8b",
+    "name": "IBM: Granite 4.1 8B",
+    "provider": "openrouter",
+    "family": "ibm-granite",
+    "created_at": "2026-04-30 19:24:31 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 131072,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.049999999999999996,
+          "output_per_million": 0.09999999999999999,
+          "cache_read_input_per_million": 0.049999999999999996
+        }
+      }
+    },
+    "metadata": {
+      "description": "Granite 4.1 8B is a dense, decoder-only 8-billion-parameter language model from IBM, part of the Granite 4.1 family. It supports a 131K-token context window and is designed for enterprise tasks...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 131072,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "inception/mercury-2",
     "name": "Mercury 2",
     "provider": "openrouter",
     "family": "mercury",
-    "created_at": "2026-03-04 00:00:00 +0530",
+    "created_at": "2026-03-04 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 50000,
     "knowledge_cutoff": null,
@@ -16945,18 +37609,48 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.25,
           "output_per_million": 0.75,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       }
     },
     "metadata": {
+      "description": "Mercury 2 is an extremely fast reasoning LLM, and the first reasoning diffusion LLM (dLLM). Instead of generating tokens sequentially, Mercury 2 produces and refines multiple tokens in parallel, achieving...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 50000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -16979,7 +37673,7 @@
     "name": "Mercury Edit 2",
     "provider": "openrouter",
     "family": null,
-    "created_at": "2026-03-30 00:00:00 +0530",
+    "created_at": "2026-03-30 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -16999,7 +37693,7 @@
         "standard": {
           "input_per_million": 0.25,
           "output_per_million": 0.75,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       }
     },
@@ -17021,12 +37715,380 @@
       }
     }
   },
+  {
+    "id": "inclusionai/ling-2.6-1t:free",
+    "name": "inclusionAI: Ling-2.6-1T (free)",
+    "provider": "openrouter",
+    "family": "inclusionai",
+    "created_at": "2026-04-23 12:43:58 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {},
+    "metadata": {
+      "description": "Ling-2.6-1T is an instant (instruct) model from inclusionAI and the company’s trillion-parameter flagship, designed for real-world agents that require fast execution and high efficiency at scale. It uses a “fast...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "inclusionai/ling-2.6-flash",
+    "name": "inclusionAI: Ling-2.6-flash",
+    "provider": "openrouter",
+    "family": "inclusionai",
+    "created_at": "2026-04-21 18:24:46 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.08,
+          "output_per_million": 0.24,
+          "cache_read_input_per_million": 0.016
+        }
+      }
+    },
+    "metadata": {
+      "description": "Ling-2.6-flash is an instant (instruct) model from inclusionAI with 104B total parameters and 7.4B active parameters, designed for real-world agents that require fast responses, strong execution, and high token efficiency....",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "inflection/inflection-3-pi",
+    "name": "Inflection: Inflection 3 Pi",
+    "provider": "openrouter",
+    "family": "inflection",
+    "created_at": "2024-10-11 00:00:00 UTC",
+    "context_window": 8000,
+    "max_output_tokens": 1024,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "Inflection 3 Pi powers Inflection's [Pi](https://pi.ai) chatbot, including backstory, emotional intelligence, productivity, and safety. It has access to recent news, and excels in scenarios like customer support and roleplay. Pi...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 8000,
+        "max_completion_tokens": 1024,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "stop",
+        "temperature",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "inflection/inflection-3-productivity",
+    "name": "Inflection: Inflection 3 Productivity",
+    "provider": "openrouter",
+    "family": "inflection",
+    "created_at": "2024-10-11 00:00:00 UTC",
+    "context_window": 8000,
+    "max_output_tokens": 1024,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "Inflection 3 Productivity is optimized for following instructions. It is better for tasks requiring JSON output or precise adherence to provided guidelines. It has access to recent news. For emotional...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 8000,
+        "max_completion_tokens": 1024,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "stop",
+        "temperature",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "kwaipilot/kat-coder-pro-v2",
+    "name": "Kwaipilot: KAT-Coder-Pro V2",
+    "provider": "openrouter",
+    "family": "kwaipilot",
+    "created_at": "2026-03-27 22:08:30 UTC",
+    "context_window": 256000,
+    "max_output_tokens": 80000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.3,
+          "output_per_million": 1.2,
+          "cache_read_input_per_million": 0.06
+        }
+      }
+    },
+    "metadata": {
+      "description": "KAT-Coder-Pro V2 is the latest high-performance model in KwaiKAT’s KAT-Coder series, designed for complex enterprise-grade software engineering and SaaS integration. It builds on the agentic coding strengths of earlier versions,...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 256000,
+        "max_completion_tokens": 80000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "liquid/lfm-2-24b-a2b",
+    "name": "LiquidAI: LFM2-24B-A2B",
+    "provider": "openrouter",
+    "family": "liquid",
+    "created_at": "2026-02-25 19:45:11 UTC",
+    "context_window": 32768,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.03,
+          "output_per_million": 0.12
+        }
+      }
+    },
+    "metadata": {
+      "description": "LFM2-24B-A2B is the largest model in the LFM2 family of hybrid architectures designed for efficient on-device deployment. Built as a 24B parameter Mixture-of-Experts model with only 2B active parameters per...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "liquid/lfm-2.5-1.2b-instruct:free",
     "name": "LFM2.5-1.2B-Instruct (free)",
     "provider": "openrouter",
     "family": "liquid",
-    "created_at": "2026-01-20 00:00:00 +0530",
+    "created_at": "2026-01-20 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -17038,9 +38100,41 @@
         "text"
       ]
     },
-    "capabilities": [],
+    "capabilities": [
+      "streaming"
+    ],
     "pricing": {},
     "metadata": {
+      "description": "LFM2.5-1.2B-Instruct is a compact, high-performance instruction-tuned model built for fast on-device AI. It delivers strong chat quality in a 1.2B parameter footprint, with efficient edge inference and broad runtime support.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -17063,7 +38157,7 @@
     "name": "LFM2.5-1.2B-Thinking (free)",
     "provider": "openrouter",
     "family": "liquid",
-    "created_at": "2026-01-20 00:00:00 +0530",
+    "created_at": "2026-01-20 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -17076,10 +38170,43 @@
       ]
     },
     "capabilities": [
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {},
     "metadata": {
+      "description": "LFM2.5-1.2B-Thinking is a lightweight reasoning-focused model optimized for agentic tasks, data extraction, and RAG—while still running comfortably on edge devices. It supports long context (up to 32K tokens) and is...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -17097,12 +38224,348 @@
       "knowledge": "2025-06"
     }
   },
+  {
+    "id": "mancer/weaver",
+    "name": "Mancer: Weaver (alpha)",
+    "provider": "openrouter",
+    "family": "mancer",
+    "created_at": "2023-08-02 00:00:00 UTC",
+    "context_window": 8000,
+    "max_output_tokens": 2000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.75,
+          "output_per_million": 1.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "An attempt to recreate Claude-style verbosity, but don't expect the same level of coherence or memory. Meant for use in roleplay/narrative situations.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama2",
+        "instruct_type": "alpaca"
+      },
+      "top_provider": {
+        "context_length": 8000,
+        "max_completion_tokens": 2000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "top_a",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "meta-llama/llama-3-70b-instruct",
+    "name": "Meta: Llama 3 70B Instruct",
+    "provider": "openrouter",
+    "family": "meta-llama",
+    "created_at": "2024-04-18 00:00:00 UTC",
+    "context_window": 8192,
+    "max_output_tokens": 8000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.51,
+          "output_per_million": 0.74
+        }
+      }
+    },
+    "metadata": {
+      "description": "Meta's latest class of model (Llama 3) launched with a variety of sizes & flavors. This 70B instruct-tuned version was optimized for high quality dialogue usecases. It has demonstrated strong...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "llama3"
+      },
+      "top_provider": {
+        "context_length": 8192,
+        "max_completion_tokens": 8000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "meta-llama/llama-3-8b-instruct",
+    "name": "Meta: Llama 3 8B Instruct",
+    "provider": "openrouter",
+    "family": "meta-llama",
+    "created_at": "2024-04-18 00:00:00 UTC",
+    "context_window": 8192,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.03,
+          "output_per_million": 0.04
+        }
+      }
+    },
+    "metadata": {
+      "description": "Meta's latest class of model (Llama 3) launched with a variety of sizes & flavors. This 8B instruct-tuned version was optimized for high quality dialogue usecases. It has demonstrated strong...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "llama3"
+      },
+      "top_provider": {
+        "context_length": 8192,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "meta-llama/llama-3.1-70b-instruct",
+    "name": "Meta: Llama 3.1 70B Instruct",
+    "provider": "openrouter",
+    "family": "meta-llama",
+    "created_at": "2024-07-23 00:00:00 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.39999999999999997,
+          "output_per_million": 0.39999999999999997
+        }
+      }
+    },
+    "metadata": {
+      "description": "Meta's latest class of model (Llama 3.1) launched with a variety of sizes & flavors. This 70B instruct-tuned version is optimized for high quality dialogue usecases. It has demonstrated strong...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "llama3"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "meta-llama/llama-3.1-8b-instruct",
+    "name": "Meta: Llama 3.1 8B Instruct",
+    "provider": "openrouter",
+    "family": "meta-llama",
+    "created_at": "2024-07-23 00:00:00 UTC",
+    "context_window": 16384,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.02,
+          "output_per_million": 0.049999999999999996
+        }
+      }
+    },
+    "metadata": {
+      "description": "Meta's latest class of model (Llama 3.1) launched with a variety of sizes & flavors. This 8B instruct-tuned version is fast and efficient. It has demonstrated strong performance compared to...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "llama3"
+      },
+      "top_provider": {
+        "context_length": 16384,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "meta-llama/llama-3.2-11b-vision-instruct",
     "name": "Llama 3.2 11B Vision Instruct",
     "provider": "openrouter",
     "family": "llama",
-    "created_at": "2024-09-25 00:00:00 +0530",
+    "created_at": "2024-09-25 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -17116,10 +38579,53 @@
       ]
     },
     "capabilities": [
-      "vision"
+      "vision",
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
     ],
-    "pricing": {},
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.245,
+          "output_per_million": 0.245
+        }
+      }
+    },
     "metadata": {
+      "description": "Llama 3.2 11B Vision is a multimodal model with 11 billion parameters, designed to handle tasks combining visual and textual data. It excels in tasks such as image captioning and...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "llama3"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -17137,12 +38643,130 @@
       "knowledge": "2023-12"
     }
   },
+  {
+    "id": "meta-llama/llama-3.2-1b-instruct",
+    "name": "Meta: Llama 3.2 1B Instruct",
+    "provider": "openrouter",
+    "family": "meta-llama",
+    "created_at": "2024-09-25 00:00:00 UTC",
+    "context_window": 60000,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.027,
+          "output_per_million": 0.19999999999999998
+        }
+      }
+    },
+    "metadata": {
+      "description": "Llama 3.2 1B is a 1-billion-parameter language model focused on efficiently performing natural language tasks, such as summarization, dialogue, and multilingual text analysis. Its smaller size allows it to operate...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "llama3"
+      },
+      "top_provider": {
+        "context_length": 60000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "meta-llama/llama-3.2-3b-instruct",
+    "name": "Meta: Llama 3.2 3B Instruct",
+    "provider": "openrouter",
+    "family": "meta-llama",
+    "created_at": "2024-09-25 00:00:00 UTC",
+    "context_window": 80000,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.051,
+          "output_per_million": 0.33999999999999997
+        }
+      }
+    },
+    "metadata": {
+      "description": "Llama 3.2 3B is a 3-billion-parameter multilingual large language model, optimized for advanced natural language processing tasks like dialogue generation, reasoning, and summarization. Designed with the latest transformer architecture, it...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "llama3"
+      },
+      "top_provider": {
+        "context_length": 80000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "meta-llama/llama-3.2-3b-instruct:free",
     "name": "Llama 3.2 3B Instruct (free)",
     "provider": "openrouter",
     "family": "llama",
-    "created_at": "2024-09-25 00:00:00 +0530",
+    "created_at": "2024-09-25 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -17156,10 +38780,38 @@
       ]
     },
     "capabilities": [
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {},
     "metadata": {
+      "description": "Llama 3.2 3B is a 3-billion-parameter multilingual large language model, optimized for advanced natural language processing tasks like dialogue generation, reasoning, and summarization. Designed with the latest transformer architecture, it...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "llama3"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -17177,12 +38829,81 @@
       "knowledge": "2023-12"
     }
   },
+  {
+    "id": "meta-llama/llama-3.3-70b-instruct",
+    "name": "Meta: Llama 3.3 70B Instruct",
+    "provider": "openrouter",
+    "family": "meta-llama",
+    "created_at": "2024-12-06 17:28:57 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.09999999999999999,
+          "output_per_million": 0.32
+        }
+      }
+    },
+    "metadata": {
+      "description": "The Meta Llama 3.3 multilingual large language model (LLM) is a pretrained and instruction tuned generative model in 70B (text in/text out). The Llama 3.3 instruction tuned text only model...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "llama3"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "meta-llama/llama-3.3-70b-instruct:free",
     "name": "Llama 3.3 70B Instruct (free)",
     "provider": "openrouter",
     "family": "llama",
-    "created_at": "2024-12-06 00:00:00 +0530",
+    "created_at": "2024-12-06 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -17196,10 +38917,40 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
     "pricing": {},
     "metadata": {
+      "description": "The Meta Llama 3.3 multilingual large language model (LLM) is a pretrained and instruction tuned generative model in 70B (text in/text out). The Llama 3.3 instruction tuned text only model...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "llama3"
+      },
+      "top_provider": {
+        "context_length": 65536,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -17217,12 +38968,469 @@
       "knowledge": "2024-12"
     }
   },
+  {
+    "id": "meta-llama/llama-4-maverick",
+    "name": "Meta: Llama 4 Maverick",
+    "provider": "openrouter",
+    "family": "meta-llama",
+    "created_at": "2025-04-05 19:37:02 UTC",
+    "context_window": 1048576,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.15,
+          "output_per_million": 0.6
+        }
+      }
+    },
+    "metadata": {
+      "description": "Llama 4 Maverick 17B Instruct (128E) is a high-capacity multimodal language model from Meta, built on a mixture-of-experts (MoE) architecture with 128 experts and 17 billion active parameters per forward...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama4",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "meta-llama/llama-4-scout",
+    "name": "Meta: Llama 4 Scout",
+    "provider": "openrouter",
+    "family": "meta-llama",
+    "created_at": "2025-04-05 19:31:59 UTC",
+    "context_window": 327680,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.08,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "description": "Llama 4 Scout 17B Instruct (16E) is a mixture-of-experts (MoE) language model developed by Meta, activating 17 billion parameters out of a total of 109B. It supports native multimodal input...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama4",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 327680,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "meta-llama/llama-guard-3-8b",
+    "name": "Llama Guard 3 8B",
+    "provider": "openrouter",
+    "family": "meta-llama",
+    "created_at": "2025-02-12 23:01:58 UTC",
+    "context_window": 131072,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.48,
+          "output_per_million": 0.03
+        }
+      }
+    },
+    "metadata": {
+      "description": "Llama Guard 3 is a Llama-3.1-8B pretrained model, fine-tuned for content safety classification. Similar to previous versions, it can be used to classify content in both LLM inputs (prompt classification)...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "none"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "meta-llama/llama-guard-4-12b",
+    "name": "Meta: Llama Guard 4 12B",
+    "provider": "openrouter",
+    "family": "meta-llama",
+    "created_at": "2025-04-30 01:06:33 UTC",
+    "context_window": 163840,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.18,
+          "output_per_million": 0.18
+        }
+      }
+    },
+    "metadata": {
+      "description": "Llama Guard 4 is a Llama 4 Scout-derived multimodal pretrained model, fine-tuned for content safety classification. Similar to previous versions, it can be used to classify content in both LLM...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 163840,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "microsoft/phi-4",
+    "name": "Microsoft: Phi 4",
+    "provider": "openrouter",
+    "family": "microsoft",
+    "created_at": "2025-01-10 06:17:52 UTC",
+    "context_window": 16384,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.065,
+          "output_per_million": 0.14
+        }
+      }
+    },
+    "metadata": {
+      "description": "[Microsoft Research](/microsoft) Phi-4 is designed to perform well in complex reasoning tasks and can operate efficiently in situations with limited memory or where quick responses are needed. At 14 billion...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 16384,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "microsoft/phi-4-mini-instruct",
+    "name": "Microsoft: Phi 4 Mini Instruct",
+    "provider": "openrouter",
+    "family": "microsoft",
+    "created_at": "2025-10-17 18:34:09 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.08,
+          "output_per_million": 0.35,
+          "cache_read_input_per_million": 0.08
+        }
+      }
+    },
+    "metadata": {
+      "description": "Phi-4-mini-instruct is a lightweight open model built upon synthetic data and filtered publicly available websites - with a focus on high-quality, reasoning dense data. The model belongs to the Phi-4...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 128000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "microsoft/wizardlm-2-8x22b",
+    "name": "WizardLM-2 8x22B",
+    "provider": "openrouter",
+    "family": "microsoft",
+    "created_at": "2024-04-16 00:00:00 UTC",
+    "context_window": 65535,
+    "max_output_tokens": 8000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.62,
+          "output_per_million": 0.62
+        }
+      }
+    },
+    "metadata": {
+      "description": "WizardLM-2 8x22B is Microsoft AI's most advanced Wizard model. It demonstrates highly competitive performance compared to leading proprietary models, and it consistently outperforms all existing state-of-the-art opensource models. It is...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": "vicuna"
+      },
+      "top_provider": {
+        "context_length": 65535,
+        "max_completion_tokens": 8000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "minimax/minimax-01",
     "name": "MiniMax-01",
     "provider": "openrouter",
     "family": "minimax",
-    "created_at": "2025-01-15 00:00:00 +0530",
+    "created_at": "2025-01-15 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 1000000,
     "knowledge_cutoff": null,
@@ -17238,7 +39446,8 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -17249,6 +39458,30 @@
       }
     },
     "metadata": {
+      "description": "MiniMax-01 is a combines MiniMax-Text-01 for text generation and MiniMax-VL-01 for image understanding. It has 456 billion parameters, with 45.9 billion parameters activated per inference, and can handle a context...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000192,
+        "max_completion_tokens": 1000192,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "temperature",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -17270,7 +39503,7 @@
     "name": "MiniMax M1",
     "provider": "openrouter",
     "family": "minimax",
-    "created_at": "2025-06-17 00:00:00 +0530",
+    "created_at": "2025-06-17 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 40000,
     "knowledge_cutoff": null,
@@ -17284,7 +39517,8 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -17295,6 +39529,39 @@
       }
     },
     "metadata": {
+      "description": "MiniMax-M1 is a large-scale, open-weight reasoning model designed for extended context and high-efficiency inference. It leverages a hybrid Mixture-of-Experts (MoE) architecture paired with a custom \"lightning attention\" mechanism, allowing it...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 40000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -17316,7 +39583,7 @@
     "name": "MiniMax M2",
     "provider": "openrouter",
     "family": "minimax",
-    "created_at": "2025-10-23 00:00:00 +0530",
+    "created_at": "2025-10-23 00:00:00 UTC",
     "context_window": 196600,
     "max_output_tokens": 118000,
     "knowledge_cutoff": null,
@@ -17331,18 +39598,58 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.28,
           "output_per_million": 1.15,
-          "cached_input_per_million": 0.28
+          "cache_read_input_per_million": 0.28,
+          "cache_write_input_per_million": 1.15
         }
       }
     },
     "metadata": {
+      "description": "MiniMax-M2 is a compact, high-efficiency large language model optimized for end-to-end coding and agentic workflows. With 10 billion activated parameters (230 billion total), it delivers near-frontier intelligence across general reasoning,...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 196608,
+        "max_completion_tokens": 196608,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -17364,12 +39671,67 @@
       }
     }
   },
+  {
+    "id": "minimax/minimax-m2-her",
+    "name": "MiniMax: MiniMax M2-her",
+    "provider": "openrouter",
+    "family": "minimax",
+    "created_at": "2026-01-23 14:07:19 UTC",
+    "context_window": 65536,
+    "max_output_tokens": 2048,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.3,
+          "output_per_million": 1.2,
+          "cache_read_input_per_million": 0.03
+        }
+      }
+    },
+    "metadata": {
+      "description": "MiniMax M2-her is a dialogue-first large language model built for immersive roleplay, character-driven chat, and expressive multi-turn conversations. Designed to stay consistent in tone and personality, it supports rich message...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 65536,
+        "max_completion_tokens": 2048,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "temperature",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "minimax/minimax-m2.1",
     "name": "MiniMax M2.1",
     "provider": "openrouter",
     "family": "minimax",
-    "created_at": "2025-12-23 00:00:00 +0530",
+    "created_at": "2025-12-23 00:00:00 UTC",
     "context_window": 204800,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -17384,7 +39746,9 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -17395,6 +39759,43 @@
       }
     },
     "metadata": {
+      "description": "MiniMax-M2.1 is a lightweight, state-of-the-art large language model optimized for coding, agentic workflows, and modern application development. With only 10 billion activated parameters, it delivers a major jump in real-world...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 196608,
+        "max_completion_tokens": 196608,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -17419,7 +39820,7 @@
     "name": "MiniMax M2.5",
     "provider": "openrouter",
     "family": "minimax",
-    "created_at": "2026-02-12 00:00:00 +0530",
+    "created_at": "2026-02-12 00:00:00 UTC",
     "context_window": 204800,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -17434,18 +39835,61 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.3,
           "output_per_million": 1.2,
-          "cached_input_per_million": 0.03
+          "cache_read_input_per_million": 0.03
         }
       }
     },
     "metadata": {
+      "description": "MiniMax-M2.5 is a SOTA large language model designed for real-world productivity. Trained in a diverse range of complex real-world digital working environments, M2.5 builds upon the coding expertise of M2.1...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 196608,
+        "max_completion_tokens": 131072,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "parallel_tool_calls",
+        "presence_penalty",
+        "reasoning",
+        "reasoning_effort",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -17471,7 +39915,7 @@
     "name": "MiniMax M2.5 (free)",
     "provider": "openrouter",
     "family": "minimax",
-    "created_at": "2026-02-12 00:00:00 +0530",
+    "created_at": "2026-02-12 00:00:00 UTC",
     "context_window": 204800,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -17486,10 +39930,39 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {},
     "metadata": {
+      "description": "MiniMax-M2.5 is a SOTA large language model designed for real-world productivity. Trained in a diverse range of complex real-world digital working environments, M2.5 builds upon the coding expertise of M2.1...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 196608,
+        "max_completion_tokens": 8192,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -17514,7 +39987,7 @@
     "name": "MiniMax M2.7",
     "provider": "openrouter",
     "family": "minimax",
-    "created_at": "2026-03-18 00:00:00 +0530",
+    "created_at": "2026-03-18 00:00:00 UTC",
     "context_window": 204800,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -17528,18 +40001,61 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.3,
           "output_per_million": 1.2,
-          "cached_input_per_million": 0.06
+          "cache_read_input_per_million": 0.06,
+          "cache_write_input_per_million": 0.375
         }
       }
     },
     "metadata": {
+      "description": "MiniMax-M2.7 is a next-generation large language model designed for autonomous, real-world productivity and continuous improvement. Built to actively participate in its own evolution, M2.7 integrates advanced agentic capabilities through multi-agent...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 196608,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -17563,7 +40079,7 @@
     "name": "Codestral 2508",
     "provider": "openrouter",
     "family": "codestral",
-    "created_at": "2025-08-01 00:00:00 +0530",
+    "created_at": "2025-08-01 00:00:00 UTC",
     "context_window": 256000,
     "max_output_tokens": 256000,
     "knowledge_cutoff": null,
@@ -17577,7 +40093,8 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -17588,6 +40105,37 @@
       }
     },
     "metadata": {
+      "description": "Mistral's cutting-edge language model for coding released end of July 2025. Codestral specializes in low-latency, high-frequency tasks such as fill-in-the-middle (FIM), code correction and test generation.\n\n[Blog Post](https://mistral.ai/news/codestral-25-08)",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 256000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -17610,7 +40158,7 @@
     "name": "Devstral 2 2512",
     "provider": "openrouter",
     "family": "devstral",
-    "created_at": "2025-09-12 00:00:00 +0530",
+    "created_at": "2025-09-12 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 262144,
     "knowledge_cutoff": null,
@@ -17624,7 +40172,8 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -17635,6 +40184,37 @@
       }
     },
     "metadata": {
+      "description": "Devstral 2 is a state-of-the-art open-source model by Mistral AI specializing in agentic coding. It is a 123B-parameter dense transformer model supporting a 256K context window. Devstral 2 supports exploring...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -17652,12 +40232,77 @@
       "knowledge": "2025-12"
     }
   },
+  {
+    "id": "mistralai/devstral-medium",
+    "name": "Mistral: Devstral Medium",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2025-07-10 15:28:41 UTC",
+    "context_window": 131072,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.39999999999999997,
+          "output_per_million": 2.0,
+          "cache_read_input_per_million": 0.04
+        }
+      }
+    },
+    "metadata": {
+      "description": "Devstral Medium is a high-performance code generation and agentic reasoning model developed jointly by Mistral AI and All Hands AI. Positioned as a step up from Devstral Small, it achieves...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "mistralai/devstral-medium-2507",
     "name": "Devstral Medium",
     "provider": "openrouter",
     "family": "devstral",
-    "created_at": "2025-07-10 00:00:00 +0530",
+    "created_at": "2025-07-10 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -17699,12 +40344,77 @@
       "knowledge": "2025-05"
     }
   },
+  {
+    "id": "mistralai/devstral-small",
+    "name": "Mistral: Devstral Small 1.1",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2025-07-10 15:19:11 UTC",
+    "context_window": 131072,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.09999999999999999,
+          "output_per_million": 0.3,
+          "cache_read_input_per_million": 0.01
+        }
+      }
+    },
+    "metadata": {
+      "description": "Devstral Small 1.1 is a 24B parameter open-weight language model for software engineering agents, developed by Mistral AI in collaboration with All Hands AI. Finetuned from Mistral Small 3.1 and...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "mistralai/devstral-small-2505",
     "name": "Devstral Small",
     "provider": "openrouter",
     "family": "devstral",
-    "created_at": "2025-05-07 00:00:00 +0530",
+    "created_at": "2025-05-07 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": null,
@@ -17750,7 +40460,7 @@
     "name": "Devstral Small 1.1",
     "provider": "openrouter",
     "family": "devstral",
-    "created_at": "2025-07-10 00:00:00 +0530",
+    "created_at": "2025-07-10 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -17792,12 +40502,543 @@
       "knowledge": "2025-05"
     }
   },
+  {
+    "id": "mistralai/ministral-14b-2512",
+    "name": "Mistral: Ministral 3 14B 2512",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2025-12-02 13:22:15 UTC",
+    "context_window": 262144,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.19999999999999998,
+          "output_per_million": 0.19999999999999998,
+          "cache_read_input_per_million": 0.02
+        }
+      }
+    },
+    "metadata": {
+      "description": "The largest model in the Ministral 3 family, Ministral 3 14B offers frontier capabilities and performance comparable to its larger Mistral Small 3.2 24B counterpart. A powerful and efficient language...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "mistralai/ministral-3b-2512",
+    "name": "Mistral: Ministral 3 3B 2512",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2025-12-02 13:19:20 UTC",
+    "context_window": 131072,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.09999999999999999,
+          "output_per_million": 0.09999999999999999,
+          "cache_read_input_per_million": 0.01
+        }
+      }
+    },
+    "metadata": {
+      "description": "The smallest model in the Ministral 3 family, Ministral 3 3B is a powerful, efficient tiny language model with vision capabilities.",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "mistralai/ministral-8b-2512",
+    "name": "Mistral: Ministral 3 8B 2512",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2025-12-02 13:20:54 UTC",
+    "context_window": 262144,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.15,
+          "output_per_million": 0.15,
+          "cache_read_input_per_million": 0.015
+        }
+      }
+    },
+    "metadata": {
+      "description": "A balanced model in the Ministral 3 family, Ministral 3 8B is a powerful, efficient tiny language model with vision capabilities.",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "mistralai/mistral-7b-instruct-v0.1",
+    "name": "Mistral: Mistral 7B Instruct v0.1",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2023-09-28 00:00:00 UTC",
+    "context_window": 2824,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.11,
+          "output_per_million": 0.19
+        }
+      }
+    },
+    "metadata": {
+      "description": "A 7.3B parameter model that outperforms Llama 2 13B on all benchmarks, with optimizations for speed and context length.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": "mistral"
+      },
+      "top_provider": {
+        "context_length": 2824,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "mistralai/mistral-large",
+    "name": "Mistral Large",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2024-02-26 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "output_per_million": 6.0,
+          "cache_read_input_per_million": 0.19999999999999998
+        }
+      }
+    },
+    "metadata": {
+      "description": "This is Mistral AI's flagship model, Mistral Large 2 (version `mistral-large-2407`). It's a proprietary weights-available model and excels at reasoning, code, JSON, chat, and more. Read the launch announcement [here](https://mistral.ai/news/mistral-large-2407/)....",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "mistralai/mistral-large-2407",
+    "name": "Mistral Large 2407",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2024-11-19 01:06:55 UTC",
+    "context_window": 131072,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "output_per_million": 6.0,
+          "cache_read_input_per_million": 0.19999999999999998
+        }
+      }
+    },
+    "metadata": {
+      "description": "This is Mistral AI's flagship model, Mistral Large 2 (version mistral-large-2407). It's a proprietary weights-available model and excels at reasoning, code, JSON, chat, and more. Read the launch announcement [here](https://mistral.ai/news/mistral-large-2407/)....",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "mistralai/mistral-large-2411",
+    "name": "Mistral Large 2411",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2024-11-19 01:11:25 UTC",
+    "context_window": 131072,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "output_per_million": 6.0,
+          "cache_read_input_per_million": 0.19999999999999998
+        }
+      }
+    },
+    "metadata": {
+      "description": "Mistral Large 2 2411 is an update of [Mistral Large 2](/mistralai/mistral-large) released together with [Pixtral Large 2411](/mistralai/pixtral-large-2411) It provides a significant upgrade on the previous [Mistral Large 24.07](/mistralai/mistral-large-2407), with notable...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "mistralai/mistral-large-2512",
+    "name": "Mistral: Mistral Large 3 2512",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2025-12-01 21:27:52 UTC",
+    "context_window": 262144,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5,
+          "cache_read_input_per_million": 0.049999999999999996
+        }
+      }
+    },
+    "metadata": {
+      "description": "Mistral Large 3 2512 is Mistral’s most capable model to date, featuring a sparse mixture-of-experts architecture with 41B active parameters (675B total), and released under the Apache 2.0 license.",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "mistralai/mistral-medium-3",
     "name": "Mistral Medium 3",
     "provider": "openrouter",
     "family": "mistral-medium",
-    "created_at": "2025-05-07 00:00:00 +0530",
+    "created_at": "2025-05-07 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -17813,7 +41054,8 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -17824,6 +41066,38 @@
       }
     },
     "metadata": {
+      "description": "Mistral Medium 3 is a high-performance enterprise-grade language model designed to deliver frontier-level capabilities at significantly reduced operational cost. It balances state-of-the-art reasoning and multimodal performance with 8× lower cost...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -17841,12 +41115,80 @@
       "knowledge": "2025-05"
     }
   },
+  {
+    "id": "mistralai/mistral-medium-3-5",
+    "name": "Mistral: Mistral Medium 3.5",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2026-04-30 17:33:59 UTC",
+    "context_window": 262144,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.5,
+          "output_per_million": 7.5
+        }
+      }
+    },
+    "metadata": {
+      "description": "Mistral Medium 3.5 is a dense 128B instruction-following model from Mistral AI. It supports text and image inputs with text output, and is designed for agentic workflows, coding, and complex...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "mistralai/mistral-medium-3.1",
     "name": "Mistral Medium 3.1",
     "provider": "openrouter",
     "family": "mistral-medium",
-    "created_at": "2025-08-12 00:00:00 +0530",
+    "created_at": "2025-08-12 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 262144,
     "knowledge_cutoff": null,
@@ -17862,7 +41204,8 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -17873,6 +41216,38 @@
       }
     },
     "metadata": {
+      "description": "Mistral Medium 3.1 is an updated version of Mistral Medium 3, which is a high-performance enterprise-grade language model designed to deliver frontier-level capabilities at significantly reduced operational cost. It balances...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -17890,12 +41265,214 @@
       "knowledge": "2025-05"
     }
   },
+  {
+    "id": "mistralai/mistral-nemo",
+    "name": "Mistral: Mistral Nemo",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2024-07-19 00:00:00 UTC",
+    "context_window": 131072,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.02,
+          "output_per_million": 0.03
+        }
+      }
+    },
+    "metadata": {
+      "description": "A 12B parameter model with a 128k token context length built by Mistral in collaboration with NVIDIA. The model is multilingual, supporting English, French, German, Spanish, Italian, Portuguese, Chinese, Japanese,...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": "mistral"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "mistralai/mistral-saba",
+    "name": "Mistral: Saba",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2025-02-17 14:40:39 UTC",
+    "context_window": 32768,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.19999999999999998,
+          "output_per_million": 0.6,
+          "cache_read_input_per_million": 0.02
+        }
+      }
+    },
+    "metadata": {
+      "description": "Mistral Saba is a 24B-parameter language model specifically designed for the Middle East and South Asia, delivering accurate and contextually relevant responses while maintaining efficient performance. Trained on curated regional...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "mistralai/mistral-small-24b-instruct-2501",
+    "name": "Mistral: Mistral Small 3",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2025-01-30 16:43:29 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.049999999999999996,
+          "output_per_million": 0.08
+        }
+      }
+    },
+    "metadata": {
+      "description": "Mistral Small 3 is a 24B-parameter language model optimized for low-latency performance across common AI tasks. Released under the Apache 2.0 license, it features both pre-trained and instruction-tuned versions designed...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "mistralai/mistral-small-2603",
     "name": "Mistral Small 4",
     "provider": "openrouter",
     "family": "mistral-small",
-    "created_at": "2026-03-16 00:00:00 +0530",
+    "created_at": "2026-03-16 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 262144,
     "knowledge_cutoff": null,
@@ -17911,7 +41488,9 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming",
+      "structured_output"
     ],
     "pricing": {
       "text_tokens": {
@@ -17922,6 +41501,41 @@
       }
     },
     "metadata": {
+      "description": "Mistral Small 4 is the next major release in the Mistral Small family, unifying the capabilities of several flagship Mistral models into a single system. It combines strong reasoning from...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -17944,7 +41558,7 @@
     "name": "Mistral Small 3.1 24B Instruct",
     "provider": "openrouter",
     "family": "mistral-small",
-    "created_at": "2025-03-17 00:00:00 +0530",
+    "created_at": "2025-03-17 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -17960,10 +41574,47 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision"
+      "vision",
+      "streaming"
     ],
-    "pricing": {},
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.35,
+          "output_per_million": 0.56
+        }
+      }
+    },
     "metadata": {
+      "description": "Mistral Small 3.1 24B Instruct is an upgraded variant of Mistral Small 3 (2501), featuring 24 billion parameters with advanced multimodal capabilities. It provides state-of-the-art performance in text-based reasoning and...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "temperature",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -17986,7 +41637,7 @@
     "name": "Mistral Small 3.2 24B Instruct",
     "provider": "openrouter",
     "family": "mistral-small",
-    "created_at": "2025-06-20 00:00:00 +0530",
+    "created_at": "2025-06-20 00:00:00 UTC",
     "context_window": 96000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -18002,10 +41653,55 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision"
+      "vision",
+      "streaming",
+      "predicted_outputs"
     ],
-    "pricing": {},
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.19999999999999998
+        }
+      }
+    },
     "metadata": {
+      "description": "Mistral-Small-3.2-24B-Instruct-2506 is an updated 24B parameter model from Mistral optimized for instruction following, repetition reduction, and improved function calling. Compared to the 3.1 release, version 3.2 significantly improves accuracy on...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -18023,12 +41719,280 @@
       "knowledge": "2024-10"
     }
   },
+  {
+    "id": "mistralai/mixtral-8x22b-instruct",
+    "name": "Mistral: Mixtral 8x22B Instruct",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2024-04-17 00:00:00 UTC",
+    "context_window": 65536,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "output_per_million": 6.0,
+          "cache_read_input_per_million": 0.19999999999999998
+        }
+      }
+    },
+    "metadata": {
+      "description": "Mistral's official instruct fine-tuned version of [Mixtral 8x22B](/models/mistralai/mixtral-8x22b). It uses 39B active parameters out of 141B, offering unparalleled cost efficiency for its size. Its strengths include: - strong math, coding,...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": "mistral"
+      },
+      "top_provider": {
+        "context_length": 65536,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "mistralai/mixtral-8x7b-instruct",
+    "name": "Mistral: Mixtral 8x7B Instruct",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2023-12-10 00:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.54,
+          "output_per_million": 0.54
+        }
+      }
+    },
+    "metadata": {
+      "description": "Mixtral 8x7B Instruct is a pretrained generative Sparse Mixture of Experts, by Mistral AI, for chat and instruction use. Incorporates 8 experts (feed-forward networks) for a total of 47 billion...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": "mistral"
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "mistralai/pixtral-large-2411",
+    "name": "Mistral: Pixtral Large 2411",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2024-11-19 00:49:48 UTC",
+    "context_window": 131072,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "output_per_million": 6.0,
+          "cache_read_input_per_million": 0.19999999999999998
+        }
+      }
+    },
+    "metadata": {
+      "description": "Pixtral Large is a 124B parameter, open-weight, multimodal model built on top of [Mistral Large 2](/mistralai/mistral-large-2411). The model is able to understand documents, charts and natural images. The model is...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "mistralai/voxtral-small-24b-2507",
+    "name": "Mistral: Voxtral Small 24B 2507",
+    "provider": "openrouter",
+    "family": "mistralai",
+    "created_at": "2025-10-30 14:39:04 UTC",
+    "context_window": 32000,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "audio"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.09999999999999999,
+          "output_per_million": 0.3,
+          "cache_read_input_per_million": 0.01
+        }
+      }
+    },
+    "metadata": {
+      "description": "Voxtral Small is an enhancement of Mistral Small 3, incorporating state-of-the-art audio input capabilities while retaining best-in-class text performance. It excels at speech transcription, translation and audio understanding. Input audio...",
+      "architecture": {
+        "modality": "text+audio->text",
+        "input_modalities": [
+          "text",
+          "audio"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 32000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "moonshotai/kimi-k2",
     "name": "Kimi K2",
     "provider": "openrouter",
     "family": "kimi",
-    "created_at": "2025-07-11 00:00:00 +0530",
+    "created_at": "2025-07-11 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -18041,7 +42005,8 @@
       ]
     },
     "capabilities": [
-      "function_calling"
+      "function_calling",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -18052,6 +42017,37 @@
       }
     },
     "metadata": {
+      "description": "Kimi K2 Instruct is a large-scale Mixture-of-Experts (MoE) language model developed by Moonshot AI, featuring 1 trillion total parameters with 32 billion active per forward pass. It is optimized for...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -18074,7 +42070,7 @@
     "name": "Kimi K2 Instruct 0905",
     "provider": "openrouter",
     "family": "kimi",
-    "created_at": "2025-09-05 00:00:00 +0530",
+    "created_at": "2025-09-05 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
@@ -18088,7 +42084,9 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -18099,6 +42097,41 @@
       }
     },
     "metadata": {
+      "description": "Kimi K2 0905 is the September update of [Kimi K2 0711](moonshotai/kimi-k2). It is a large-scale Mixture-of-Experts (MoE) language model developed by Moonshot AI, featuring 1 trillion total parameters with 32...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 262144,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -18121,7 +42154,7 @@
     "name": "Kimi K2 Instruct 0905 (exacto)",
     "provider": "openrouter",
     "family": "kimi",
-    "created_at": "2025-09-05 00:00:00 +0530",
+    "created_at": "2025-09-05 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
@@ -18168,7 +42201,7 @@
     "name": "Kimi K2 Thinking",
     "provider": "openrouter",
     "family": "kimi-thinking",
-    "created_at": "2025-11-06 00:00:00 +0530",
+    "created_at": "2025-11-06 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 262144,
     "knowledge_cutoff": null,
@@ -18183,18 +42216,57 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.6,
           "output_per_million": 2.5,
-          "cached_input_per_million": 0.15
+          "cache_read_input_per_million": 0.15
         }
       }
     },
     "metadata": {
+      "description": "Kimi K2 Thinking is Moonshot AI’s most advanced open reasoning model to date, extending the K2 series into agentic, long-horizon reasoning. Built on the trillion-parameter Mixture-of-Experts (MoE) architecture introduced in...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 262144,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -18221,7 +42293,7 @@
     "name": "Kimi K2.5",
     "provider": "openrouter",
     "family": "kimi",
-    "created_at": "2026-01-27 00:00:00 +0530",
+    "created_at": "2026-01-27 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 262144,
     "knowledge_cutoff": null,
@@ -18239,18 +42311,61 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.6,
           "output_per_million": 3,
-          "cached_input_per_million": 0.1
+          "cache_read_input_per_million": 0.1
         }
       }
     },
     "metadata": {
+      "description": "Kimi K2.5 is Moonshot AI's native multimodal model, delivering state-of-the-art visual coding capability and a self-directed agent swarm paradigm. Built on Kimi K2 with continued pretraining over approximately 15T mixed...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 65535,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "parallel_tool_calls",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -18272,12 +42387,409 @@
       "knowledge": "2025-01"
     }
   },
+  {
+    "id": "moonshotai/kimi-k2.6",
+    "name": "Kimi K2.6",
+    "provider": "openrouter",
+    "family": "kimi",
+    "created_at": "2026-04-20 00:00:00 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 262144,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "vision",
+      "streaming",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.95,
+          "output_per_million": 4,
+          "cache_read_input_per_million": 0.16
+        }
+      }
+    },
+    "metadata": {
+      "description": "Kimi K2.6 is Moonshot AI's next-generation multimodal model, designed for long-horizon coding, coding-driven UI/UX generation, and multi-agent orchestration. It handles complex end-to-end coding tasks across Python, Rust, and Go, and...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "parallel_tool_calls",
+        "presence_penalty",
+        "reasoning",
+        "reasoning_effort",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ],
+      "source": "models.dev",
+      "provider_id": "openrouter",
+      "open_weights": true,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2026-04-20",
+      "interleaved": {
+        "field": "reasoning_details"
+      },
+      "cost": {
+        "input": 0.95,
+        "output": 4,
+        "cache_read": 0.16
+      },
+      "limit": {
+        "context": 262144,
+        "output": 262144
+      }
+    }
+  },
+  {
+    "id": "morph/morph-v3-fast",
+    "name": "Morph: Morph V3 Fast",
+    "provider": "openrouter",
+    "family": "morph",
+    "created_at": "2025-07-07 17:40:02 UTC",
+    "context_window": 81920,
+    "max_output_tokens": 38000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.7999999999999999,
+          "output_per_million": 1.2
+        }
+      }
+    },
+    "metadata": {
+      "description": "Morph's fastest apply model for code edits. ~10,500 tokens/sec with 96% accuracy for rapid code transformations. The model requires the prompt to be in the following format: <instruction>{instruction}</instruction> <code>{initial_code}</code> <update>{edit_snippet}</update>...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 81920,
+        "max_completion_tokens": 38000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "stop",
+        "temperature"
+      ]
+    }
+  },
+  {
+    "id": "morph/morph-v3-large",
+    "name": "Morph: Morph V3 Large",
+    "provider": "openrouter",
+    "family": "morph",
+    "created_at": "2025-07-07 17:54:18 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 131072,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.8999999999999999,
+          "output_per_million": 1.9
+        }
+      }
+    },
+    "metadata": {
+      "description": "Morph's high-accuracy apply model for complex code edits. ~4,500 tokens/sec with 98% accuracy for precise code transformations. The model requires the prompt to be in the following format: <instruction>{instruction}</instruction> <code>{initial_code}</code>...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 131072,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "stop",
+        "temperature"
+      ]
+    }
+  },
+  {
+    "id": "nex-agi/deepseek-v3.1-nex-n1",
+    "name": "Nex AGI: DeepSeek V3.1 Nex N1",
+    "provider": "openrouter",
+    "family": "nex-agi",
+    "created_at": "2025-12-08 14:33:13 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 163840,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.135,
+          "output_per_million": 0.5
+        }
+      }
+    },
+    "metadata": {
+      "description": "DeepSeek V3.1 Nex-N1 is the flagship release of the Nex-N1 series — a post-trained model designed to highlight agent autonomy, tool use, and real-world productivity. Nex-N1 demonstrates competitive performance across...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "DeepSeek",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 163840,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "response_format",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "nousresearch/hermes-2-pro-llama-3-8b",
+    "name": "NousResearch: Hermes 2 Pro - Llama-3 8B",
+    "provider": "openrouter",
+    "family": "nousresearch",
+    "created_at": "2024-05-27 00:00:00 UTC",
+    "context_window": 8192,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.14,
+          "output_per_million": 0.14
+        }
+      }
+    },
+    "metadata": {
+      "description": "Hermes 2 Pro is an upgraded, retrained version of Nous Hermes 2, consisting of an updated and cleaned version of the OpenHermes 2.5 Dataset, as well as a newly introduced...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "chatml"
+      },
+      "top_provider": {
+        "context_length": 8192,
+        "max_completion_tokens": 8192,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "nousresearch/hermes-3-llama-3.1-405b",
+    "name": "Nous: Hermes 3 405B Instruct",
+    "provider": "openrouter",
+    "family": "nousresearch",
+    "created_at": "2024-08-16 00:00:00 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.0,
+          "output_per_million": 1.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "Hermes 3 is a generalist language model with many improvements over Hermes 2, including advanced agentic capabilities, much better roleplaying, reasoning, multi-turn conversation, long context coherence, and improvements across the...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "chatml"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "nousresearch/hermes-3-llama-3.1-405b:free",
     "name": "Hermes 3 405B Instruct (free)",
     "provider": "openrouter",
     "family": "hermes",
-    "created_at": "2024-08-16 00:00:00 +0530",
+    "created_at": "2024-08-16 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -18290,10 +42802,38 @@
       ]
     },
     "capabilities": [
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {},
     "metadata": {
+      "description": "Hermes 3 is a generalist language model with many improvements over Hermes 2, including advanced agentic capabilities, much better roleplaying, reasoning, multi-turn conversation, long context coherence, and improvements across the...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "chatml"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -18311,12 +42851,78 @@
       "knowledge": "2023-12"
     }
   },
+  {
+    "id": "nousresearch/hermes-3-llama-3.1-70b",
+    "name": "Nous: Hermes 3 70B Instruct",
+    "provider": "openrouter",
+    "family": "nousresearch",
+    "created_at": "2024-08-18 00:00:00 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.3,
+          "output_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "description": "Hermes 3 is a generalist language model with many improvements over [Hermes 2](/models/nousresearch/nous-hermes-2-mistral-7b-dpo), including advanced agentic capabilities, much better roleplaying, reasoning, multi-turn conversation, long context coherence, and improvements across the...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "chatml"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "nousresearch/hermes-4-405b",
     "name": "Hermes 4 405B",
     "provider": "openrouter",
     "family": "hermes",
-    "created_at": "2025-08-25 00:00:00 +0530",
+    "created_at": "2025-08-25 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -18330,7 +42936,9 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "structured_output"
     ],
     "pricing": {
       "text_tokens": {
@@ -18341,6 +42949,36 @@
       }
     },
     "metadata": {
+      "description": "Hermes 4 is a large-scale reasoning model built on Meta-Llama-3.1-405B and released by Nous Research. It introduces a hybrid reasoning mode, where the model can choose to deliberate internally with...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "temperature",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -18363,7 +43001,7 @@
     "name": "Hermes 4 70B",
     "provider": "openrouter",
     "family": "hermes",
-    "created_at": "2025-08-25 00:00:00 +0530",
+    "created_at": "2025-08-25 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -18378,7 +43016,8 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -18389,6 +43028,36 @@
       }
     },
     "metadata": {
+      "description": "Hermes 4 70B is a hybrid reasoning model from Nous Research, built on Meta-Llama-3.1-70B. It introduces the same hybrid mode as the larger 405B release, allowing the model to either...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "temperature",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -18406,12 +43075,220 @@
       "knowledge": "2023-12"
     }
   },
+  {
+    "id": "nvidia/llama-3.1-nemotron-70b-instruct",
+    "name": "NVIDIA: Llama 3.1 Nemotron 70B Instruct",
+    "provider": "openrouter",
+    "family": "nvidia",
+    "created_at": "2024-10-15 00:00:00 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.2,
+          "output_per_million": 1.2
+        }
+      }
+    },
+    "metadata": {
+      "description": "NVIDIA's Llama 3.1 Nemotron 70B is a language model designed for generating precise and useful responses. Leveraging [Llama 3.1 70B](/models/meta-llama/llama-3.1-70b-instruct) architecture and Reinforcement Learning from Human Feedback (RLHF), it excels...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "llama3"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "nvidia/llama-3.3-nemotron-super-49b-v1.5",
+    "name": "NVIDIA: Llama 3.3 Nemotron Super 49B V1.5",
+    "provider": "openrouter",
+    "family": "nvidia",
+    "created_at": "2025-10-10 13:03:15 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.09999999999999999,
+          "output_per_million": 0.39999999999999997
+        }
+      }
+    },
+    "metadata": {
+      "description": "Llama-3.3-Nemotron-Super-49B-v1.5 is a 49B-parameter, English-centric reasoning/chat model derived from Meta’s Llama-3.3-70B-Instruct with a 128K context. It’s post-trained for agentic workflows (RAG, tool calling) via SFT across math, code, science, and...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "nvidia/nemotron-3-nano-30b-a3b",
+    "name": "NVIDIA: Nemotron 3 Nano 30B A3B",
+    "provider": "openrouter",
+    "family": "nvidia",
+    "created_at": "2025-12-14 16:54:35 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 228000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.049999999999999996,
+          "output_per_million": 0.19999999999999998
+        }
+      }
+    },
+    "metadata": {
+      "description": "NVIDIA Nemotron 3 Nano 30B A3B is a small language MoE model with highest compute efficiency and accuracy for developers to build specialized agentic AI systems. The model is fully...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 228000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "nvidia/nemotron-3-nano-30b-a3b:free",
     "name": "Nemotron 3 Nano 30B A3B (free)",
     "provider": "openrouter",
     "family": "nemotron",
-    "created_at": "2025-12-14 00:00:00 +0530",
+    "created_at": "2025-12-14 00:00:00 UTC",
     "context_window": 256000,
     "max_output_tokens": 256000,
     "knowledge_cutoff": null,
@@ -18426,10 +43303,39 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {},
     "metadata": {
+      "description": "NVIDIA Nemotron 3 Nano 30B A3B is a small language MoE model with highest compute efficiency and accuracy for developers to build specialized agentic AI systems. The model is fully...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 256000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "seed",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -18447,12 +43353,88 @@
       "knowledge": "2025-11"
     }
   },
+  {
+    "id": "nvidia/nemotron-3-nano-omni-30b-a3b-reasoning:free",
+    "name": "Nemotron 3 Nano Omni (free)",
+    "provider": "openrouter",
+    "family": "nemotron",
+    "created_at": "2026-04-28 00:00:00 UTC",
+    "context_window": 256000,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video",
+        "audio"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {},
+    "metadata": {
+      "description": "NVIDIA Nemotron™ 3 Nano Omni is a 30B-A3B open multimodal model designed to function as a perception and context sub-agent in enterprise agent systems. It accepts text, image, video, and...",
+      "architecture": {
+        "modality": "text+image+audio+video->text",
+        "input_modalities": [
+          "text",
+          "audio",
+          "image",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 256000,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "seed",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
+      "source": "models.dev",
+      "provider_id": "openrouter",
+      "open_weights": true,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2026-04-28",
+      "cost": {
+        "input": 0,
+        "output": 0
+      },
+      "limit": {
+        "context": 256000,
+        "output": 65536
+      }
+    }
+  },
   {
     "id": "nvidia/nemotron-3-super-120b-a12b",
     "name": "Nemotron 3 Super",
     "provider": "openrouter",
     "family": "nemotron",
-    "created_at": "2026-03-11 00:00:00 +0530",
+    "created_at": "2026-03-11 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 262144,
     "knowledge_cutoff": null,
@@ -18466,7 +43448,10 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -18477,6 +43462,44 @@
       }
     },
     "metadata": {
+      "description": "NVIDIA Nemotron 3 Super is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications. Built on a hybrid Mamba-Transformer...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -18499,7 +43522,7 @@
     "name": "Nemotron 3 Super (free)",
     "provider": "openrouter",
     "family": "nemotron",
-    "created_at": "2026-03-11 00:00:00 +0530",
+    "created_at": "2026-03-11 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 262144,
     "knowledge_cutoff": null,
@@ -18513,10 +43536,42 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "structured_output"
     ],
     "pricing": {},
     "metadata": {
+      "description": "NVIDIA Nemotron 3 Super is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications. Built on a hybrid Mamba-Transformer...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 262144,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -18534,12 +43589,83 @@
       "knowledge": "2024-04"
     }
   },
+  {
+    "id": "nvidia/nemotron-nano-12b-v2-vl",
+    "name": "NVIDIA: Nemotron Nano 12B 2 VL",
+    "provider": "openrouter",
+    "family": "nvidia",
+    "created_at": "2025-10-28 18:19:25 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.19999999999999998,
+          "output_per_million": 0.6
+        }
+      }
+    },
+    "metadata": {
+      "description": "NVIDIA Nemotron Nano 2 VL is a 12-billion-parameter open multimodal reasoning model designed for video understanding and document intelligence. It introduces a hybrid Transformer-Mamba architecture, combining transformer-level accuracy with Mamba’s...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "nvidia/nemotron-nano-12b-v2-vl:free",
     "name": "Nemotron Nano 12B 2 VL (free)",
     "provider": "openrouter",
     "family": "nemotron",
-    "created_at": "2025-10-28 00:00:00 +0530",
+    "created_at": "2025-10-28 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": null,
@@ -18555,10 +43681,41 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {},
     "metadata": {
+      "description": "NVIDIA Nemotron Nano 2 VL is a 12-billion-parameter open multimodal reasoning model designed for video understanding and document intelligence. It introduces a hybrid Transformer-Mamba architecture, combining transformer-level accuracy with Mamba’s...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 128000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "seed",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -18581,7 +43738,7 @@
     "name": "nvidia-nemotron-nano-9b-v2",
     "provider": "openrouter",
     "family": "nemotron",
-    "created_at": "2025-08-18 00:00:00 +0530",
+    "created_at": "2025-08-18 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -18595,7 +43752,10 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -18606,6 +43766,42 @@
       }
     },
     "metadata": {
+      "description": "NVIDIA-Nemotron-Nano-9B-v2 is a large language model (LLM) trained from scratch by NVIDIA, and designed as a unified model for both reasoning and non-reasoning tasks. It responds to user queries and...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -18628,7 +43824,7 @@
     "name": "Nemotron Nano 9B V2 (free)",
     "provider": "openrouter",
     "family": "nemotron",
-    "created_at": "2025-09-05 00:00:00 +0530",
+    "created_at": "2025-09-05 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": null,
@@ -18643,10 +43839,41 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {},
     "metadata": {
+      "description": "NVIDIA-Nemotron-Nano-9B-v2 is a large language model (LLM) trained from scratch by NVIDIA, and designed as a unified model for both reasoning and non-reasoning tasks. It responds to user queries and...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -18664,12 +43891,616 @@
       "knowledge": "2024-09"
     }
   },
+  {
+    "id": "openai/gpt-3.5-turbo",
+    "name": "OpenAI: GPT-3.5 Turbo",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2023-05-28 00:00:00 UTC",
+    "context_window": 16385,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "description": "GPT-3.5 Turbo is OpenAI's fastest model. It can understand and generate natural language or code, and is optimized for chat and traditional completion tasks.\n\nTraining data up to Sep 2021.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 16385,
+        "max_completion_tokens": 4096,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "openai/gpt-3.5-turbo-0613",
+    "name": "OpenAI: GPT-3.5 Turbo (older v0613)",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2024-01-25 00:00:00 UTC",
+    "context_window": 4095,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.0,
+          "output_per_million": 2.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "GPT-3.5 Turbo is OpenAI's fastest model. It can understand and generate natural language or code, and is optimized for chat and traditional completion tasks.\n\nTraining data up to Sep 2021.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 4095,
+        "max_completion_tokens": 4096,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_completion_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "openai/gpt-3.5-turbo-16k",
+    "name": "OpenAI: GPT-3.5 Turbo 16k",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2023-08-28 00:00:00 UTC",
+    "context_window": 16385,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 3.0,
+          "output_per_million": 4.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "This model offers four times the context length of gpt-3.5-turbo, allowing it to support approximately 20 pages of text in a single request at a higher cost. Training data: up...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 16385,
+        "max_completion_tokens": 4096,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_completion_tokens",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "openai/gpt-3.5-turbo-instruct",
+    "name": "OpenAI: GPT-3.5 Turbo Instruct",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2023-09-28 00:00:00 UTC",
+    "context_window": 4095,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.5,
+          "output_per_million": 2.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "This model is a variant of GPT-3.5 Turbo tuned for instructional prompts and omitting chat-related optimizations. Training data: up to Sep 2021.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": "chatml"
+      },
+      "top_provider": {
+        "context_length": 4095,
+        "max_completion_tokens": 4096,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "openai/gpt-4",
+    "name": "OpenAI: GPT-4",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2023-05-28 00:00:00 UTC",
+    "context_window": 8191,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 30.0,
+          "output_per_million": 60.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "OpenAI's flagship model, GPT-4 is a large-scale multimodal language model capable of solving difficult problems with greater accuracy than previous models due to its broader general knowledge and advanced reasoning...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 8191,
+        "max_completion_tokens": 4096,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_completion_tokens",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "openai/gpt-4-0314",
+    "name": "OpenAI: GPT-4 (older v0314)",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2023-05-28 00:00:00 UTC",
+    "context_window": 8191,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 30.0,
+          "output_per_million": 60.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "GPT-4-0314 is the first version of GPT-4 released, with a context length of 8,192 tokens, and was supported until June 14. Training data: up to Sep 2021.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 8191,
+        "max_completion_tokens": 4096,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "openai/gpt-4-1106-preview",
+    "name": "OpenAI: GPT-4 Turbo (older v1106)",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2023-11-06 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 10.0,
+          "output_per_million": 30.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "The latest GPT-4 Turbo model with vision capabilities. Vision requests can now use JSON mode and function calling.\n\nTraining data: up to April 2023.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 4096,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "openai/gpt-4-turbo",
+    "name": "OpenAI: GPT-4 Turbo",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2024-04-09 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 10.0,
+          "output_per_million": 30.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "The latest GPT-4 Turbo model with vision capabilities. Vision requests can now use JSON mode and function calling.\n\nTraining data: up to December 2023.",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 4096,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "openai/gpt-4-turbo-preview",
+    "name": "OpenAI: GPT-4 Turbo Preview",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2024-01-25 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 10.0,
+          "output_per_million": 30.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "The preview GPT-4 model with improved instruction following, JSON mode, reproducible outputs, parallel function calling, and more. Training data: up to Dec 2023. **Note:** heavily rate limited by OpenAI while...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 4096,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "openai/gpt-4.1",
     "name": "GPT-4.1",
     "provider": "openrouter",
     "family": "gpt",
-    "created_at": "2025-04-14 00:00:00 +0530",
+    "created_at": "2025-04-14 00:00:00 UTC",
     "context_window": 1047576,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -18685,18 +44516,50 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 2,
           "output_per_million": 8,
-          "cached_input_per_million": 0.5
+          "cache_read_input_per_million": 0.5
         }
       }
     },
     "metadata": {
+      "description": "GPT-4.1 is a flagship large language model optimized for advanced instruction following, real-world software engineering, and long-context reasoning. It supports a 1 million token context window and outperforms GPT-4o and...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1047576,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_completion_tokens",
+        "max_tokens",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -18720,7 +44583,7 @@
     "name": "GPT-4.1 Mini",
     "provider": "openrouter",
     "family": "gpt-mini",
-    "created_at": "2025-04-14 00:00:00 +0530",
+    "created_at": "2025-04-14 00:00:00 UTC",
     "context_window": 1047576,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -18736,18 +44599,50 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.4,
           "output_per_million": 1.6,
-          "cached_input_per_million": 0.1
+          "cache_read_input_per_million": 0.1
         }
       }
     },
     "metadata": {
+      "description": "GPT-4.1 Mini is a mid-sized model delivering performance competitive with GPT-4o at substantially lower latency and cost. It retains a 1 million token context window and scores 45.1% on hard...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1047576,
+        "max_completion_tokens": 32768,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_completion_tokens",
+        "max_tokens",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -18766,12 +44661,443 @@
       "knowledge": "2024-04"
     }
   },
+  {
+    "id": "openai/gpt-4.1-nano",
+    "name": "OpenAI: GPT-4.1 Nano",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2025-04-14 17:22:49 UTC",
+    "context_window": 1047576,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.09999999999999999,
+          "output_per_million": 0.39999999999999997,
+          "cache_read_input_per_million": 0.024999999999999998
+        }
+      }
+    },
+    "metadata": {
+      "description": "For tasks that demand low latency, GPT‑4.1 nano is the fastest and cheapest model in the GPT-4.1 series. It delivers exceptional performance at a small size with its 1 million...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1047576,
+        "max_completion_tokens": 32768,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_completion_tokens",
+        "max_tokens",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "openai/gpt-4o",
+    "name": "OpenAI: GPT-4o",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2024-05-13 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "GPT-4o (\"o\" for \"omni\") is OpenAI's latest AI model, supporting both text and image inputs with text outputs. It maintains the intelligence level of [GPT-4 Turbo](/models/openai/gpt-4-turbo) while being twice as...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_completion_tokens",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p",
+        "web_search_options"
+      ]
+    }
+  },
+  {
+    "id": "openai/gpt-4o-2024-05-13",
+    "name": "OpenAI: GPT-4o (2024-05-13)",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2024-05-13 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5.0,
+          "output_per_million": 15.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "GPT-4o (\"o\" for \"omni\") is OpenAI's latest AI model, supporting both text and image inputs with text outputs. It maintains the intelligence level of [GPT-4 Turbo](/models/openai/gpt-4-turbo) while being twice as...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 4096,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_completion_tokens",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p",
+        "web_search_options"
+      ]
+    }
+  },
+  {
+    "id": "openai/gpt-4o-2024-08-06",
+    "name": "OpenAI: GPT-4o (2024-08-06)",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2024-08-06 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 1.25
+        }
+      }
+    },
+    "metadata": {
+      "description": "The 2024-08-06 version of GPT-4o offers improved performance in structured outputs, with the ability to supply a JSON schema in the respone_format. Read more [here](https://openai.com/index/introducing-structured-outputs-in-the-api/). GPT-4o (\"o\" for \"omni\") is...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_completion_tokens",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p",
+        "web_search_options"
+      ]
+    }
+  },
+  {
+    "id": "openai/gpt-4o-2024-11-20",
+    "name": "OpenAI: GPT-4o (2024-11-20)",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2024-11-20 18:33:14 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0,
+          "cache_read_input_per_million": 1.25
+        }
+      }
+    },
+    "metadata": {
+      "description": "The 2024-11-20 version of GPT-4o offers a leveled-up creative writing ability with more natural, engaging, and tailored writing to improve relevance & readability. It’s also better at working with uploaded...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 16384,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p",
+        "web_search_options"
+      ]
+    }
+  },
+  {
+    "id": "openai/gpt-4o-audio-preview",
+    "name": "OpenAI: GPT-4o Audio",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2025-08-15 04:44:21 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "audio",
+        "text"
+      ],
+      "output": [
+        "text",
+        "audio"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "The gpt-4o-audio-preview model adds support for audio inputs as prompts. This enhancement allows the model to detect nuances within audio recordings and add depth to generated user experiences. Audio outputs...",
+      "architecture": {
+        "modality": "text+audio->text+audio",
+        "input_modalities": [
+          "audio",
+          "text"
+        ],
+        "output_modalities": [
+          "text",
+          "audio"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 16384,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "openai/gpt-4o-mini",
     "name": "GPT-4o-mini",
     "provider": "openrouter",
     "family": "gpt-mini",
-    "created_at": "2024-07-18 00:00:00 +0530",
+    "created_at": "2024-07-18 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
@@ -18787,18 +45113,57 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.15,
           "output_per_million": 0.6,
-          "cached_input_per_million": 0.08
+          "cache_read_input_per_million": 0.08
         }
       }
     },
     "metadata": {
+      "description": "GPT-4o mini is OpenAI's newest model after [GPT-4 Omni](/models/openai/gpt-4o), supporting both text and image inputs with text outputs. As their most advanced small model, it is many multiples more affordable...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_completion_tokens",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p",
+        "web_search_options"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -18817,12 +45182,197 @@
       "knowledge": "2024-10"
     }
   },
+  {
+    "id": "openai/gpt-4o-mini-2024-07-18",
+    "name": "OpenAI: GPT-4o-mini (2024-07-18)",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2024-07-18 00:00:00 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.15,
+          "output_per_million": 0.6,
+          "cache_read_input_per_million": 0.075
+        }
+      }
+    },
+    "metadata": {
+      "description": "GPT-4o mini is OpenAI's newest model after [GPT-4 Omni](/models/openai/gpt-4o), supporting both text and image inputs with text outputs. As their most advanced small model, it is many multiples more affordable...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 16384,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p",
+        "web_search_options"
+      ]
+    }
+  },
+  {
+    "id": "openai/gpt-4o-mini-search-preview",
+    "name": "OpenAI: GPT-4o-mini Search Preview",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2025-03-12 22:22:02 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.15,
+          "output_per_million": 0.6
+        }
+      }
+    },
+    "metadata": {
+      "description": "GPT-4o mini Search Preview is a specialized model for web search in Chat Completions. It is trained to understand and execute web search queries.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 16384,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "response_format",
+        "structured_outputs",
+        "web_search_options"
+      ]
+    }
+  },
+  {
+    "id": "openai/gpt-4o-search-preview",
+    "name": "OpenAI: GPT-4o Search Preview",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2025-03-12 22:19:09 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "GPT-4o Search Previewis a specialized model for web search in Chat Completions. It is trained to understand and execute web search queries.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 16384,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "response_format",
+        "structured_outputs",
+        "web_search_options"
+      ]
+    }
+  },
   {
     "id": "openai/gpt-5",
     "name": "GPT-5",
     "provider": "openrouter",
     "family": "gpt",
-    "created_at": "2025-08-07 00:00:00 +0530",
+    "created_at": "2025-08-07 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-10-01",
@@ -18839,7 +45389,8 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -18850,6 +45401,37 @@
       }
     },
     "metadata": {
+      "description": "GPT-5 is OpenAI’s most advanced model, offering major improvements in reasoning, code quality, and user experience. It is optimized for complex tasks that require step-by-step reasoning, instruction following, and accuracy...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -18872,7 +45454,7 @@
     "name": "GPT-5 Chat (latest)",
     "provider": "openrouter",
     "family": "gpt-codex",
-    "created_at": "2025-08-07 00:00:00 +0530",
+    "created_at": "2025-08-07 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-09-30",
@@ -18888,7 +45470,8 @@
     "capabilities": [
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -18899,6 +45482,32 @@
       }
     },
     "metadata": {
+      "description": "GPT-5 Chat is designed for advanced, natural, multimodal, and context-aware conversations for enterprise applications.",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "file",
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 16384,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "response_format",
+        "seed",
+        "structured_outputs"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -18921,7 +45530,7 @@
     "name": "GPT-5 Codex",
     "provider": "openrouter",
     "family": "gpt-codex",
-    "created_at": "2025-09-15 00:00:00 +0530",
+    "created_at": "2025-09-15 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-10-01",
@@ -18938,18 +45547,48 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
         }
       }
     },
     "metadata": {
+      "description": "GPT-5-Codex is a specialized version of GPT-5 optimized for software engineering and coding workflows. It is designed for both interactive development sessions and long, independent execution of complex engineering tasks....",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -18973,7 +45612,7 @@
     "name": "GPT-5 Image",
     "provider": "openrouter",
     "family": "gpt",
-    "created_at": "2025-10-14 00:00:00 +0530",
+    "created_at": "2025-10-14 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-10-01",
@@ -18992,18 +45631,56 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 5,
           "output_per_million": 10,
-          "cached_input_per_million": 1.25
+          "cache_read_input_per_million": 1.25
         }
       }
     },
     "metadata": {
+      "description": "[GPT-5](https://openrouter.ai/openai/gpt-5) Image combines OpenAI's GPT-5 model with state-of-the-art image generation capabilities. It offers major improvements in reasoning, code quality, and user experience while incorporating GPT Image 1's superior instruction following,...",
+      "architecture": {
+        "modality": "text+image+file->text+image",
+        "input_modalities": [
+          "image",
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "image",
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19022,12 +45699,85 @@
       "knowledge": "2024-10-01"
     }
   },
+  {
+    "id": "openai/gpt-5-image-mini",
+    "name": "OpenAI: GPT-5 Image Mini",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2025-10-16 14:23:03 UTC",
+    "context_window": 400000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "file",
+        "image",
+        "text"
+      ],
+      "output": [
+        "image",
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 2.0,
+          "cache_read_input_per_million": 0.25
+        }
+      }
+    },
+    "metadata": {
+      "description": "GPT-5 Image Mini combines OpenAI's advanced language capabilities, powered by [GPT-5 Mini](https://openrouter.ai/openai/gpt-5-mini), with GPT Image 1 Mini for efficient image generation. This natively multimodal model features superior instruction following, text...",
+      "architecture": {
+        "modality": "text+image+file->text+image",
+        "input_modalities": [
+          "file",
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "image",
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "openai/gpt-5-mini",
     "name": "GPT-5 Mini",
     "provider": "openrouter",
     "family": "gpt-mini",
-    "created_at": "2025-08-07 00:00:00 +0530",
+    "created_at": "2025-08-07 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-10-01",
@@ -19044,7 +45794,8 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -19055,6 +45806,37 @@
       }
     },
     "metadata": {
+      "description": "GPT-5 Mini is a compact version of GPT-5, designed to handle lighter-weight reasoning tasks. It provides the same instruction-following and safety-tuning benefits as GPT-5, but with reduced latency and cost....",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19077,7 +45859,7 @@
     "name": "GPT-5 Nano",
     "provider": "openrouter",
     "family": "gpt-nano",
-    "created_at": "2025-08-07 00:00:00 +0530",
+    "created_at": "2025-08-07 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-10-01",
@@ -19094,7 +45876,8 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -19105,6 +45888,37 @@
       }
     },
     "metadata": {
+      "description": "GPT-5-Nano is the smallest and fastest variant in the GPT-5 system, optimized for developer tools, rapid interactions, and ultra-low latency environments. While limited in reasoning depth compared to its larger...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19127,7 +45941,7 @@
     "name": "GPT-5 Pro",
     "provider": "openrouter",
     "family": "gpt-pro",
-    "created_at": "2025-10-06 00:00:00 +0530",
+    "created_at": "2025-10-06 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 272000,
     "knowledge_cutoff": "2024-09-30",
@@ -19144,7 +45958,8 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -19155,6 +45970,36 @@
       }
     },
     "metadata": {
+      "description": "GPT-5 Pro is OpenAI’s most advanced model, offering major improvements in reasoning, code quality, and user experience. It is optimized for complex tasks that require step-by-step reasoning, instruction following, and...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19177,7 +46022,7 @@
     "name": "GPT-5.1",
     "provider": "openrouter",
     "family": "gpt",
-    "created_at": "2025-11-13 00:00:00 +0530",
+    "created_at": "2025-11-13 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-09-30",
@@ -19194,18 +46039,50 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
         }
       }
     },
     "metadata": {
+      "description": "GPT-5.1 is the latest frontier-grade model in the GPT-5 series, offering stronger general-purpose reasoning, improved instruction adherence, and a more natural conversational style compared to GPT-5. It uses adaptive reasoning...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19229,7 +46106,7 @@
     "name": "GPT-5.1 Chat",
     "provider": "openrouter",
     "family": "gpt-codex",
-    "created_at": "2025-11-13 00:00:00 +0530",
+    "created_at": "2025-11-13 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": "2024-09-30",
@@ -19246,18 +46123,48 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
         }
       }
     },
     "metadata": {
+      "description": "GPT-5.1 Chat (AKA Instant is the fast, lightweight member of the 5.1 family, optimized for low-latency chat while retaining strong general intelligence. It uses adaptive reasoning to selectively “think” on...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "file",
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 16384,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_completion_tokens",
+        "max_tokens",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19281,7 +46188,7 @@
     "name": "GPT-5.1-Codex",
     "provider": "openrouter",
     "family": "gpt-codex",
-    "created_at": "2025-11-13 00:00:00 +0530",
+    "created_at": "2025-11-13 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-09-30",
@@ -19298,18 +46205,49 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.125
+          "cache_read_input_per_million": 0.125
         }
       }
     },
     "metadata": {
+      "description": "GPT-5.1-Codex is a specialized version of GPT-5.1 optimized for software engineering and coding workflows. It is designed for both interactive development sessions and long, independent execution of complex engineering tasks....",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19333,7 +46271,7 @@
     "name": "GPT-5.1-Codex-Max",
     "provider": "openrouter",
     "family": "gpt-codex",
-    "created_at": "2025-11-13 00:00:00 +0530",
+    "created_at": "2025-11-13 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2024-09-30",
@@ -19350,18 +46288,49 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.1,
           "output_per_million": 9,
-          "cached_input_per_million": 0.11
+          "cache_read_input_per_million": 0.11
         }
       }
     },
     "metadata": {
+      "description": "GPT-5.1-Codex-Max is OpenAI’s latest agentic coding model, designed for long-running, high-context software development tasks. It is based on an updated version of the 5.1 reasoning stack and trained on agentic...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19385,7 +46354,7 @@
     "name": "GPT-5.1-Codex-Mini",
     "provider": "openrouter",
     "family": "gpt-codex",
-    "created_at": "2025-11-13 00:00:00 +0530",
+    "created_at": "2025-11-13 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 100000,
     "knowledge_cutoff": "2024-09-30",
@@ -19402,18 +46371,49 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.25,
           "output_per_million": 2,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       }
     },
     "metadata": {
+      "description": "GPT-5.1-Codex-Mini is a smaller and faster version of GPT-5.1-Codex",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19437,7 +46437,7 @@
     "name": "GPT-5.2",
     "provider": "openrouter",
     "family": "gpt",
-    "created_at": "2025-12-11 00:00:00 +0530",
+    "created_at": "2025-12-11 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2025-08-31",
@@ -19454,18 +46454,50 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.75,
           "output_per_million": 14,
-          "cached_input_per_million": 0.175
+          "cache_read_input_per_million": 0.175
         }
       }
     },
     "metadata": {
+      "description": "GPT-5.2 is the latest frontier-grade model in the GPT-5 series, offering stronger agentic and long context perfomance compared to GPT-5.1. It uses adaptive reasoning to allocate computation dynamically, responding quickly...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "file",
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19489,7 +46521,7 @@
     "name": "GPT-5.2 Chat",
     "provider": "openrouter",
     "family": "gpt-codex",
-    "created_at": "2025-12-11 00:00:00 +0530",
+    "created_at": "2025-12-11 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": "2025-08-31",
@@ -19506,18 +46538,48 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.75,
           "output_per_million": 14,
-          "cached_input_per_million": 0.175
+          "cache_read_input_per_million": 0.175
         }
       }
     },
     "metadata": {
+      "description": "GPT-5.2 Chat (AKA Instant) is the fast, lightweight member of the 5.2 family, optimized for low-latency chat while retaining strong general intelligence. It uses adaptive reasoning to selectively “think” on...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "file",
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 32000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_completion_tokens",
+        "max_tokens",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19541,7 +46603,7 @@
     "name": "GPT-5.2-Codex",
     "provider": "openrouter",
     "family": "gpt-codex",
-    "created_at": "2026-01-14 00:00:00 +0530",
+    "created_at": "2026-01-14 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2025-08-31",
@@ -19558,18 +46620,49 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.75,
           "output_per_million": 14,
-          "cached_input_per_million": 0.175
+          "cache_read_input_per_million": 0.175
         }
       }
     },
     "metadata": {
+      "description": "GPT-5.2-Codex is an upgraded version of GPT-5.1-Codex optimized for software engineering and coding workflows. It is designed for both interactive development sessions and long, independent execution of complex engineering tasks....",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19593,7 +46686,7 @@
     "name": "GPT-5.2 Pro",
     "provider": "openrouter",
     "family": "gpt-pro",
-    "created_at": "2025-12-11 00:00:00 +0530",
+    "created_at": "2025-12-11 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2025-08-31",
@@ -19610,7 +46703,8 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -19621,6 +46715,36 @@
       }
     },
     "metadata": {
+      "description": "GPT-5.2 Pro is OpenAI’s most advanced model, offering major improvements in agentic coding and long context performance over GPT-5 Pro. It is optimized for complex tasks that require step-by-step reasoning,...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19638,12 +46762,77 @@
       "knowledge": "2025-08-31"
     }
   },
+  {
+    "id": "openai/gpt-5.3-chat",
+    "name": "OpenAI: GPT-5.3 Chat",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2026-03-03 18:54:21 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.75,
+          "output_per_million": 14.0,
+          "cache_read_input_per_million": 0.175
+        }
+      }
+    },
+    "metadata": {
+      "description": "GPT-5.3 Chat is an update to ChatGPT's most-used model that makes everyday conversations smoother, more useful, and more directly helpful. It delivers more accurate answers with better contextualization and significantly...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 16384,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_completion_tokens",
+        "max_tokens",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ]
+    }
+  },
   {
     "id": "openai/gpt-5.3-codex",
     "name": "GPT-5.3-Codex",
     "provider": "openrouter",
     "family": "gpt-codex",
-    "created_at": "2026-02-24 00:00:00 +0530",
+    "created_at": "2026-02-24 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2025-08-31",
@@ -19661,18 +46850,50 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.75,
           "output_per_million": 14,
-          "cached_input_per_million": 0.175
+          "cache_read_input_per_million": 0.175
         }
       }
     },
     "metadata": {
+      "description": "GPT-5.3-Codex is OpenAI’s most advanced agentic coding model, combining the frontier software engineering performance of GPT-5.2-Codex with the broader reasoning and professional knowledge capabilities of GPT-5.2. It achieves state-of-the-art results...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19696,7 +46917,7 @@
     "name": "GPT-5.4",
     "provider": "openrouter",
     "family": "gpt",
-    "created_at": "2026-03-05 00:00:00 +0530",
+    "created_at": "2026-03-05 00:00:00 UTC",
     "context_window": 1050000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2025-08-31",
@@ -19714,18 +46935,50 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 2.5,
           "output_per_million": 15,
-          "cached_input_per_million": 0.25
+          "cache_read_input_per_million": 0.25
         }
       }
     },
     "metadata": {
+      "description": "GPT-5.4 is OpenAI’s latest frontier model, unifying the Codex and GPT lines into a single system. It features a 1M+ token context window (922K input, 128K output) with support for...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1050000,
+        "max_completion_tokens": 128000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19750,12 +47003,83 @@
       "knowledge": "2025-08-31"
     }
   },
+  {
+    "id": "openai/gpt-5.4-image-2",
+    "name": "OpenAI: GPT-5.4 Image 2",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2026-04-21 18:52:08 UTC",
+    "context_window": 272000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text",
+        "file"
+      ],
+      "output": [
+        "image",
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 8.0,
+          "output_per_million": 15.0,
+          "cache_read_input_per_million": 2.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "[GPT-5.4](https://openrouter.ai/openai/gpt-5.4) Image 2 combines OpenAI's GPT-5.4 model with state-of-the-art image generation capabilities from GPT Image 2. It enables rich multimodal workflows, allowing users to seamlessly move between reasoning, coding, and...",
+      "architecture": {
+        "modality": "text+image+file->text+image",
+        "input_modalities": [
+          "image",
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "image",
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 272000,
+        "max_completion_tokens": 128000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "top_logprobs"
+      ]
+    }
+  },
   {
     "id": "openai/gpt-5.4-mini",
     "name": "GPT-5.4 Mini",
     "provider": "openrouter",
     "family": "gpt-mini",
-    "created_at": "2026-03-17 00:00:00 +0530",
+    "created_at": "2026-03-17 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2025-08-31",
@@ -19773,18 +47097,50 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
-          "input_per_million": 0.00000075,
-          "output_per_million": 0.0000045,
-          "cached_input_per_million": 0.000000075
+          "input_per_million": 0.75,
+          "output_per_million": 4.5,
+          "cache_read_input_per_million": 0.075
         }
       }
     },
     "metadata": {
+      "description": "GPT-5.4 mini brings the core capabilities of GPT-5.4 to a faster, more efficient model optimized for high-throughput workloads. It supports text and image inputs with strong performance across reasoning, coding,...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "file",
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19792,9 +47148,9 @@
       "temperature": true,
       "last_updated": "2026-03-17",
       "cost": {
-        "input": 0.00000075,
-        "output": 0.0000045,
-        "cache_read": 0.000000075
+        "input": 0.75,
+        "output": 4.5,
+        "cache_read": 0.075
       },
       "limit": {
         "context": 400000,
@@ -19808,7 +47164,7 @@
     "name": "GPT-5.4 Nano",
     "provider": "openrouter",
     "family": "gpt-nano",
-    "created_at": "2026-03-17 00:00:00 +0530",
+    "created_at": "2026-03-17 00:00:00 UTC",
     "context_window": 400000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2025-08-31",
@@ -19825,18 +47181,50 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
-          "input_per_million": 0.0000002,
-          "output_per_million": 0.00000125,
-          "cached_input_per_million": 0.00000002
+          "input_per_million": 0.2,
+          "output_per_million": 1.25,
+          "cache_read_input_per_million": 0.02
         }
       }
     },
     "metadata": {
+      "description": "GPT-5.4 nano is the most lightweight and cost-efficient variant of the GPT-5.4 family, optimized for speed-critical and high-volume tasks. It supports text and image inputs and is designed for low-latency...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "file",
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19844,9 +47232,9 @@
       "temperature": true,
       "last_updated": "2026-03-17",
       "cost": {
-        "input": 0.0000002,
-        "output": 0.00000125,
-        "cache_read": 0.00000002
+        "input": 0.2,
+        "output": 1.25,
+        "cache_read": 0.02
       },
       "limit": {
         "context": 400000,
@@ -19860,7 +47248,7 @@
     "name": "GPT-5.4 Pro",
     "provider": "openrouter",
     "family": "gpt-pro",
-    "created_at": "2026-03-05 00:00:00 +0530",
+    "created_at": "2026-03-05 00:00:00 UTC",
     "context_window": 1050000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": "2025-08-31",
@@ -19877,18 +47265,51 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming",
+      "structured_output"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 30,
           "output_per_million": 180,
-          "cached_input_per_million": 30
+          "cache_read_input_per_million": 30
         }
       }
     },
     "metadata": {
+      "description": "GPT-5.4 Pro is OpenAI's most advanced model, building on GPT-5.4's unified architecture with enhanced reasoning capabilities for complex, high-stakes tasks. It features a 1M+ token context window (922K input, 128K...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1050000,
+        "max_completion_tokens": 128000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -19908,12 +47329,402 @@
       "knowledge": "2025-08-31"
     }
   },
+  {
+    "id": "openai/gpt-5.5",
+    "name": "GPT-5.5",
+    "provider": "openrouter",
+    "family": "gpt",
+    "created_at": "2026-04-23 00:00:00 UTC",
+    "context_window": 1050000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": "2025-12-01",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5,
+          "output_per_million": 30,
+          "cache_read_input_per_million": 0.5
+        }
+      }
+    },
+    "metadata": {
+      "description": "GPT-5.5 is OpenAI’s frontier model designed for complex professional workloads, building on GPT-5.4 with stronger reasoning, higher reliability, and improved token efficiency on hard tasks. It features a 1M+ token...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "file",
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1050000,
+        "max_completion_tokens": 128000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
+      "source": "models.dev",
+      "provider_id": "openrouter",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": false,
+      "last_updated": "2026-04-23",
+      "cost": {
+        "input": 5,
+        "output": 30,
+        "cache_read": 0.5,
+        "context_over_200k": {
+          "input": 10,
+          "output": 45,
+          "cache_read": 1
+        }
+      },
+      "limit": {
+        "context": 1050000,
+        "input": 922000,
+        "output": 128000
+      },
+      "knowledge": "2025-12-01"
+    }
+  },
+  {
+    "id": "openai/gpt-5.5-pro",
+    "name": "GPT-5.5 Pro",
+    "provider": "openrouter",
+    "family": "gpt-pro",
+    "created_at": "2026-04-23 00:00:00 UTC",
+    "context_window": 1050000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": "2025-12-01",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 30,
+          "output_per_million": 180
+        }
+      }
+    },
+    "metadata": {
+      "description": "GPT-5.5 Pro is OpenAI’s high-capability model optimized for deep reasoning and accuracy on complex, high-stakes workloads. It features a 1M+ token context window (922K input, 128K output) with support for...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "file",
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1050000,
+        "max_completion_tokens": 128000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
+      "source": "models.dev",
+      "provider_id": "openrouter",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": false,
+      "last_updated": "2026-04-23",
+      "cost": {
+        "input": 30,
+        "output": 180,
+        "context_over_200k": {
+          "input": 60,
+          "output": 270
+        }
+      },
+      "limit": {
+        "context": 1050000,
+        "input": 922000,
+        "output": 128000
+      },
+      "knowledge": "2025-12-01"
+    }
+  },
+  {
+    "id": "openai/gpt-audio",
+    "name": "OpenAI: GPT Audio",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2026-01-19 22:42:49 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "audio"
+      ],
+      "output": [
+        "text",
+        "audio"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.5,
+          "output_per_million": 10.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "The gpt-audio model is OpenAI's first generally available audio model. The new snapshot features an upgraded decoder for more natural sounding voices and maintains better voice consistency. Audio is priced...",
+      "architecture": {
+        "modality": "text+audio->text+audio",
+        "input_modalities": [
+          "text",
+          "audio"
+        ],
+        "output_modalities": [
+          "text",
+          "audio"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 16384,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "openai/gpt-audio-mini",
+    "name": "OpenAI: GPT Audio Mini",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2026-01-19 21:50:19 UTC",
+    "context_window": 128000,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "audio"
+      ],
+      "output": [
+        "text",
+        "audio"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.6,
+          "output_per_million": 2.4
+        }
+      }
+    },
+    "metadata": {
+      "description": "A cost-efficient version of GPT Audio. The new snapshot features an upgraded decoder for more natural sounding voices and maintains better voice consistency. Input is priced at $0.60 per million...",
+      "architecture": {
+        "modality": "text+audio->text+audio",
+        "input_modalities": [
+          "text",
+          "audio"
+        ],
+        "output_modalities": [
+          "text",
+          "audio"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": 16384,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "openai/gpt-chat-latest",
+    "name": "OpenAI: GPT Chat Latest",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2026-05-05 16:56:52 UTC",
+    "context_window": 400000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5.0,
+          "output_per_million": 30.0,
+          "cache_read_input_per_million": 0.5
+        }
+      }
+    },
+    "metadata": {
+      "description": "GPT Chat Latest points to OpenAI's stable API alias `chat-latest` that always resolves to the latest Instant chat model used in ChatGPT. As OpenAI rolls out new Instant model updates...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "tool_choice",
+        "tools",
+        "top_logprobs"
+      ]
+    }
+  },
   {
     "id": "openai/gpt-oss-120b",
     "name": "GPT OSS 120B",
     "provider": "openrouter",
     "family": "gpt-oss",
-    "created_at": "2025-08-05 00:00:00 +0530",
+    "created_at": "2025-08-05 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -19928,7 +47739,9 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -19939,6 +47752,45 @@
       }
     },
     "metadata": {
+      "description": "gpt-oss-120b is an open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning, agentic, and general-purpose production use cases. It activates 5.1B parameters per forward pass and is optimized...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -19960,7 +47812,7 @@
     "name": "GPT OSS 120B (exacto)",
     "provider": "openrouter",
     "family": "gpt-oss",
-    "created_at": "2025-08-05 00:00:00 +0530",
+    "created_at": "2025-08-05 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -20007,7 +47859,7 @@
     "name": "gpt-oss-120b (free)",
     "provider": "openrouter",
     "family": "gpt-oss",
-    "created_at": "2025-08-05 00:00:00 +0530",
+    "created_at": "2025-08-05 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -20021,10 +47873,39 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {},
     "metadata": {
+      "description": "gpt-oss-120b is an open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning, agentic, and general-purpose production use cases. It activates 5.1B parameters per forward pass and is optimized...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 131072,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -20046,7 +47927,7 @@
     "name": "GPT OSS 20B",
     "provider": "openrouter",
     "family": "gpt-oss",
-    "created_at": "2025-08-05 00:00:00 +0530",
+    "created_at": "2025-08-05 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -20061,7 +47942,9 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -20072,6 +47955,45 @@
       }
     },
     "metadata": {
+      "description": "gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license. It uses a Mixture-of-Experts (MoE) architecture with 3.6B active parameters per forward pass, optimized for...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 131072,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -20093,7 +48015,7 @@
     "name": "gpt-oss-20b (free)",
     "provider": "openrouter",
     "family": "gpt-oss",
-    "created_at": "2025-08-05 00:00:00 +0530",
+    "created_at": "2025-08-05 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -20107,10 +48029,39 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {},
     "metadata": {
+      "description": "gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license. It uses a Mixture-of-Experts (MoE) architecture with 3.6B active parameters per forward pass, optimized for...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 8192,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -20132,7 +48083,7 @@
     "name": "GPT OSS Safeguard 20B",
     "provider": "openrouter",
     "family": "gpt-oss",
-    "created_at": "2025-10-29 00:00:00 +0530",
+    "created_at": "2025-10-29 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -20146,7 +48097,9 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "structured_output"
     ],
     "pricing": {
       "text_tokens": {
@@ -20157,6 +48110,36 @@
       }
     },
     "metadata": {
+      "description": "gpt-oss-safeguard-20b is a safety reasoning model from OpenAI built upon gpt-oss-20b. This open-weight, 21B-parameter Mixture-of-Experts (MoE) model offers lower latency for safety tasks like content classification, LLM filtering, and trust...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -20173,12 +48156,473 @@
       }
     }
   },
+  {
+    "id": "openai/o1",
+    "name": "OpenAI: o1",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2024-12-17 18:26:39 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 100000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 15.0,
+          "output_per_million": 60.0,
+          "cache_read_input_per_million": 7.5
+        }
+      }
+    },
+    "metadata": {
+      "description": "The latest and strongest model family from OpenAI, o1 is designed to spend more time thinking before responding. The o1 model series is trained with large-scale reinforcement learning to reason...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 100000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ]
+    }
+  },
+  {
+    "id": "openai/o1-pro",
+    "name": "OpenAI: o1-pro",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2025-03-19 22:26:51 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 100000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 150.0,
+          "output_per_million": 600.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "The o1 series of models are trained with reinforcement learning to think before they answer and perform complex reasoning. The o1-pro model uses more compute to think harder and provide...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 100000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs"
+      ]
+    }
+  },
+  {
+    "id": "openai/o3",
+    "name": "OpenAI: o3",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2025-04-16 17:10:57 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 100000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "output_per_million": 8.0,
+          "cache_read_input_per_million": 0.5
+        }
+      }
+    },
+    "metadata": {
+      "description": "o3 is a well-rounded and powerful model across domains. It sets a new standard for math, science, coding, and visual reasoning tasks. It also excels at technical writing and instruction-following....",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 100000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ]
+    }
+  },
+  {
+    "id": "openai/o3-deep-research",
+    "name": "OpenAI: o3 Deep Research",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2025-10-10 20:54:21 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 100000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 10.0,
+          "output_per_million": 40.0,
+          "cache_read_input_per_million": 2.5
+        }
+      }
+    },
+    "metadata": {
+      "description": "o3-deep-research is OpenAI's advanced model for deep research, designed to tackle complex, multi-step research tasks.\n\nNote: This model always uses the 'web_search' tool which adds additional cost.",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 100000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "openai/o3-mini",
+    "name": "OpenAI: o3 Mini",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2025-01-31 19:28:41 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 100000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.1,
+          "output_per_million": 4.4,
+          "cache_read_input_per_million": 0.55
+        }
+      }
+    },
+    "metadata": {
+      "description": "OpenAI o3-mini is a cost-efficient language model optimized for STEM reasoning tasks, particularly excelling in science, mathematics, and coding. This model supports the `reasoning_effort` parameter, which can be set to...",
+      "architecture": {
+        "modality": "text+file->text",
+        "input_modalities": [
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 100000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ]
+    }
+  },
+  {
+    "id": "openai/o3-mini-high",
+    "name": "OpenAI: o3 Mini High",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2025-02-12 15:03:31 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 100000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.1,
+          "output_per_million": 4.4,
+          "cache_read_input_per_million": 0.55
+        }
+      }
+    },
+    "metadata": {
+      "description": "OpenAI o3-mini-high is the same model as [o3-mini](/openai/o3-mini) with reasoning_effort set to high. o3-mini is a cost-efficient language model optimized for STEM reasoning tasks, particularly excelling in science, mathematics, and...",
+      "architecture": {
+        "modality": "text+file->text",
+        "input_modalities": [
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 100000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ]
+    }
+  },
+  {
+    "id": "openai/o3-pro",
+    "name": "OpenAI: o3 Pro",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2025-06-10 23:32:32 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 100000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "file",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 20.0,
+          "output_per_million": 80.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "The o-series of models are trained with reinforcement learning to think before they answer and perform complex reasoning. The o3-pro model uses more compute to think harder and provide consistently...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "file",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 100000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ]
+    }
+  },
   {
     "id": "openai/o4-mini",
     "name": "o4 Mini",
     "provider": "openrouter",
     "family": "o-mini",
-    "created_at": "2025-04-16 00:00:00 +0530",
+    "created_at": "2025-04-16 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 100000,
     "knowledge_cutoff": null,
@@ -20195,18 +48639,49 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.1,
           "output_per_million": 4.4,
-          "cached_input_per_million": 0.28
+          "cache_read_input_per_million": 0.28
         }
       }
     },
     "metadata": {
+      "description": "OpenAI o4-mini is a compact reasoning model in the o-series, optimized for fast, cost-efficient performance while retaining strong multimodal and agentic capabilities. It supports tool use and demonstrates competitive reasoning...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 100000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -20225,12 +48700,273 @@
       "knowledge": "2024-06"
     }
   },
+  {
+    "id": "openai/o4-mini-deep-research",
+    "name": "OpenAI: o4 Mini Deep Research",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2025-10-10 20:54:02 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 100000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "file",
+        "image",
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "output_per_million": 8.0,
+          "cache_read_input_per_million": 0.5
+        }
+      }
+    },
+    "metadata": {
+      "description": "o4-mini-deep-research is OpenAI's faster, more affordable deep research model—ideal for tackling complex, multi-step research tasks.\n\nNote: This model always uses the 'web_search' tool which adds additional cost.",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "file",
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 100000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "openai/o4-mini-high",
+    "name": "OpenAI: o4 Mini High",
+    "provider": "openrouter",
+    "family": "openai",
+    "created_at": "2025-04-16 17:23:32 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 100000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.1,
+          "output_per_million": 4.4,
+          "cache_read_input_per_million": 0.275
+        }
+      }
+    },
+    "metadata": {
+      "description": "OpenAI o4-mini-high is the same model as [o4-mini](/openai/o4-mini) with reasoning_effort set to high. OpenAI o4-mini is a compact reasoning model in the o-series, optimized for fast, cost-efficient performance while retaining...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "GPT",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 100000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ]
+    }
+  },
+  {
+    "id": "openrouter/auto",
+    "name": "Auto Router",
+    "provider": "openrouter",
+    "family": "openrouter",
+    "created_at": "2023-11-08 00:00:00 UTC",
+    "context_window": 2000000,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "audio",
+        "file",
+        "video"
+      ],
+      "output": [
+        "text",
+        "image"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {},
+    "metadata": {
+      "description": "\"Your prompt will be processed by a meta-model and routed to one of dozens of models (see below), optimizing for the best possible output. To see which model was used,...",
+      "architecture": {
+        "modality": "text+image+file+audio+video->text+image",
+        "input_modalities": [
+          "text",
+          "image",
+          "audio",
+          "file",
+          "video"
+        ],
+        "output_modalities": [
+          "text",
+          "image"
+        ],
+        "tokenizer": "Router",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": null,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_completion_tokens",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p",
+        "web_search_options"
+      ]
+    }
+  },
+  {
+    "id": "openrouter/bodybuilder",
+    "name": "Body Builder (beta)",
+    "provider": "openrouter",
+    "family": "openrouter",
+    "created_at": "2025-12-05 03:00:53 UTC",
+    "context_window": 128000,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {},
+    "metadata": {
+      "description": "Transform your natural language requests into structured OpenRouter API request objects. Describe what you want to accomplish with AI models, and Body Builder will construct the appropriate API calls. Example:...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Router",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": null,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": []
+    }
+  },
   {
     "id": "openrouter/elephant-alpha",
     "name": "Elephant (free)",
     "provider": "openrouter",
     "family": "elephant",
-    "created_at": "2026-04-13 00:00:00 +0530",
+    "created_at": "2026-04-13 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -20269,8 +49005,8 @@
     "id": "openrouter/free",
     "name": "Free Models Router",
     "provider": "openrouter",
-    "family": null,
-    "created_at": "2026-02-01 00:00:00 +0530",
+    "family": "openrouter",
+    "created_at": "2026-02-01 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 8000,
     "knowledge_cutoff": null,
@@ -20287,10 +49023,48 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {},
     "metadata": {
+      "description": "The simplest way to get free inference. openrouter/free is a router that selects free models at random from the models available on OpenRouter. The router smartly filters for models that...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Router",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": null,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -20308,12 +49082,592 @@
       }
     }
   },
+  {
+    "id": "openrouter/owl-alpha",
+    "name": "Owl Alpha",
+    "provider": "openrouter",
+    "family": "openrouter",
+    "created_at": "2026-04-28 00:00:00 UTC",
+    "context_window": 1048756,
+    "max_output_tokens": 262144,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
+    ],
+    "pricing": {},
+    "metadata": {
+      "description": "Owl Alpha is a high-performance foundation model designed for agentic workloads. Natively supports tool use, and long-context tasks, with strong performance in code generation, automated workflows, and complex instruction execution....",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048756,
+        "max_completion_tokens": 262144,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
+      "source": "models.dev",
+      "provider_id": "openrouter",
+      "open_weights": false,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2026-04-30",
+      "status": "alpha",
+      "cost": {
+        "input": 0,
+        "output": 0
+      },
+      "limit": {
+        "context": 1048756,
+        "output": 262144
+      }
+    }
+  },
+  {
+    "id": "openrouter/pareto-code",
+    "name": "Pareto Code Router",
+    "provider": "openrouter",
+    "family": "openrouter",
+    "created_at": "2026-04-21 00:00:00 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 200000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "streaming"
+    ],
+    "pricing": {},
+    "metadata": {
+      "description": "The Pareto Router is a way to have OpenRouter always pick a strong coding model for your needs without committing to a specific one. You express a single `min_coding_score` preference...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Router",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": null,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [],
+      "source": "models.dev",
+      "provider_id": "openrouter",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2026-04-21",
+      "limit": {
+        "context": 200000,
+        "output": 200000
+      }
+    }
+  },
+  {
+    "id": "perplexity/sonar",
+    "name": "Perplexity: Sonar",
+    "provider": "openrouter",
+    "family": "perplexity",
+    "created_at": "2025-01-27 21:36:48 UTC",
+    "context_window": 127072,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.0,
+          "output_per_million": 1.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "Sonar is lightweight, affordable, fast, and simple to use — now featuring citations and the ability to customize sources. It is designed for companies seeking to integrate lightweight question-and-answer features...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 127072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "temperature",
+        "top_k",
+        "top_p",
+        "web_search_options"
+      ]
+    }
+  },
+  {
+    "id": "perplexity/sonar-deep-research",
+    "name": "Perplexity: Sonar Deep Research",
+    "provider": "openrouter",
+    "family": "perplexity",
+    "created_at": "2025-03-07 01:34:06 UTC",
+    "context_window": 128000,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "output_per_million": 8.0,
+          "reasoning_output_per_million": 3.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "Sonar Deep Research is a research-focused model designed for multi-step retrieval, synthesis, and reasoning across complex topics. It autonomously searches, reads, and evaluates sources, refining its approach as it gathers...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": "deepseek-r1"
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "temperature",
+        "top_k",
+        "top_p",
+        "web_search_options"
+      ]
+    }
+  },
+  {
+    "id": "perplexity/sonar-pro",
+    "name": "Perplexity: Sonar Pro",
+    "provider": "openrouter",
+    "family": "perplexity",
+    "created_at": "2025-03-07 01:53:43 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 8000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 3.0,
+          "output_per_million": 15.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "Note: Sonar Pro pricing includes Perplexity search pricing. See [details here](https://docs.perplexity.ai/guides/pricing#detailed-pricing-breakdown-for-sonar-reasoning-pro-and-sonar-pro) For enterprises seeking more advanced capabilities, the Sonar Pro API can handle in-depth, multi-step queries with added extensibility, like...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 8000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "temperature",
+        "top_k",
+        "top_p",
+        "web_search_options"
+      ]
+    }
+  },
+  {
+    "id": "perplexity/sonar-pro-search",
+    "name": "Perplexity: Sonar Pro Search",
+    "provider": "openrouter",
+    "family": "perplexity",
+    "created_at": "2025-10-30 19:59:26 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 8000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 3.0,
+          "output_per_million": 15.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "Exclusively available on the OpenRouter API, Sonar Pro's new Pro Search mode is Perplexity's most advanced agentic search system. It is designed for deeper reasoning and analysis. Pricing is based...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 8000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_p",
+        "web_search_options"
+      ]
+    }
+  },
+  {
+    "id": "perplexity/sonar-reasoning-pro",
+    "name": "Perplexity: Sonar Reasoning Pro",
+    "provider": "openrouter",
+    "family": "perplexity",
+    "created_at": "2025-03-07 02:08:28 UTC",
+    "context_window": 128000,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "output_per_million": 8.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "Note: Sonar Pro pricing includes Perplexity search pricing. See [details here](https://docs.perplexity.ai/guides/pricing#detailed-pricing-breakdown-for-sonar-reasoning-pro-and-sonar-pro) Sonar Reasoning Pro is a premier reasoning model powered by DeepSeek R1 with Chain of Thought (CoT). Designed for...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": "deepseek-r1"
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "temperature",
+        "top_k",
+        "top_p",
+        "web_search_options"
+      ]
+    }
+  },
+  {
+    "id": "poolside/laguna-m.1:free",
+    "name": "Laguna M.1",
+    "provider": "openrouter",
+    "family": "poolside",
+    "created_at": "2026-04-28 00:00:00 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "streaming"
+    ],
+    "pricing": {},
+    "metadata": {
+      "description": "Laguna M.1 is the flagship coding agent model from [Poolside](https://poolside.ai), optimized for complex software engineering tasks. Designed for agentic coding workflows, it supports tool calling and reasoning, with a 128K...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 8192,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "temperature",
+        "tool_choice",
+        "tools"
+      ],
+      "source": "models.dev",
+      "provider_id": "openrouter",
+      "open_weights": false,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2026-04-28",
+      "interleaved": {
+        "field": "reasoning_content"
+      },
+      "cost": {
+        "input": 0,
+        "output": 0,
+        "cache_read": 0,
+        "cache_write": 0
+      },
+      "limit": {
+        "context": 131072,
+        "output": 8192
+      }
+    }
+  },
+  {
+    "id": "poolside/laguna-xs.2:free",
+    "name": "Laguna XS.2",
+    "provider": "openrouter",
+    "family": "poolside",
+    "created_at": "2026-04-28 00:00:00 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "streaming"
+    ],
+    "pricing": {},
+    "metadata": {
+      "description": "Laguna XS.2 is the second-generation model in the XS size class from [Poolside](https://poolside.ai), their efficient coding agent series. It combines tool calling and reasoning capabilities with a compact footprint, offering...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 8192,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "temperature",
+        "tool_choice",
+        "tools"
+      ],
+      "source": "models.dev",
+      "provider_id": "openrouter",
+      "open_weights": true,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2026-04-28",
+      "interleaved": {
+        "field": "reasoning_content"
+      },
+      "cost": {
+        "input": 0,
+        "output": 0,
+        "cache_read": 0,
+        "cache_write": 0
+      },
+      "limit": {
+        "context": 131072,
+        "output": 8192
+      }
+    }
+  },
   {
     "id": "prime-intellect/intellect-3",
     "name": "Intellect 3",
     "provider": "openrouter",
     "family": "glm",
-    "created_at": "2025-01-15 00:00:00 +0530",
+    "created_at": "2025-01-15 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -20328,7 +49682,8 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -20339,6 +49694,39 @@
       }
     },
     "metadata": {
+      "description": "INTELLECT-3 is a 106B-parameter Mixture-of-Experts model (12B active) post-trained from GLM-4.5-Air-Base using supervised fine-tuning (SFT) followed by large-scale reinforcement learning (RL). It offers state-of-the-art performance for its size across math,...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 131072,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -20356,12 +49744,149 @@
       "knowledge": "2024-10"
     }
   },
+  {
+    "id": "qwen/qwen-2.5-72b-instruct",
+    "name": "Qwen2.5 72B Instruct",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2024-09-19 00:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.36,
+          "output_per_million": 0.39999999999999997
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen2.5 72B is the latest series of Qwen large language models. Qwen2.5 brings the following improvements upon Qwen2: - Significantly more knowledge and has greatly improved capabilities in coding and...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen",
+        "instruct_type": "chatml"
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen-2.5-7b-instruct",
+    "name": "Qwen: Qwen2.5 7B Instruct",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2024-10-16 00:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.04,
+          "output_per_million": 0.09999999999999999
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen2.5 7B is the latest series of Qwen large language models. Qwen2.5 brings the following improvements upon Qwen2: - Significantly more knowledge and has greatly improved capabilities in coding and...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen",
+        "instruct_type": "chatml"
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "qwen/qwen-2.5-coder-32b-instruct",
     "name": "Qwen2.5 Coder 32B Instruct",
     "provider": "openrouter",
     "family": "qwen",
-    "created_at": "2024-11-11 00:00:00 +0530",
+    "created_at": "2024-11-11 00:00:00 UTC",
     "context_window": 32768,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -20374,10 +49899,46 @@
       ]
     },
     "capabilities": [
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
-    "pricing": {},
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.66,
+          "output_per_million": 1.0
+        }
+      }
+    },
     "metadata": {
+      "description": "Qwen2.5-Coder is the latest series of Code-Specific Qwen large language models (formerly known as CodeQwen). Qwen2.5-Coder brings the following improvements upon CodeQwen1.5: - Significantly improvements in **code generation**, **code reasoning**...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen",
+        "instruct_type": "chatml"
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "temperature",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -20395,12 +49956,498 @@
       "knowledge": "2024-10"
     }
   },
+  {
+    "id": "qwen/qwen-3.6-27b",
+    "name": "Qwen3.6 27B",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2026-04-22 00:00:00 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 81920,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.195,
+          "output_per_million": 1.56
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "openrouter",
+      "open_weights": true,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2026-04-22",
+      "cost": {
+        "input": 0.195,
+        "output": 1.56
+      },
+      "limit": {
+        "context": 262144,
+        "output": 81920
+      },
+      "knowledge": "2025-04"
+    }
+  },
+  {
+    "id": "qwen/qwen-max",
+    "name": "Qwen: Qwen-Max ",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-02-01 09:31:29 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.04,
+          "output_per_million": 4.16,
+          "cache_read_input_per_million": 0.20800000000000002
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen-Max, based on Qwen2.5, provides the best inference performance among [Qwen models](/qwen), especially for complex multi-step tasks. It's a large-scale MoE model that has been pretrained on over 20 trillion...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": 8192,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen-plus",
+    "name": "Qwen: Qwen-Plus",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-02-01 11:37:20 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.26,
+          "output_per_million": 0.78,
+          "cache_read_input_per_million": 0.052000000000000005
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen-Plus, based on the Qwen2.5 foundation model, is a 131K context model with a balanced performance, speed, and cost combination.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen-plus-2025-07-28",
+    "name": "Qwen: Qwen Plus 0728",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-09-08 16:06:39 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.26,
+          "output_per_million": 0.78
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen Plus 0728, based on the Qwen3 foundation model, is a 1 million context hybrid reasoning model with a balanced performance, speed, and cost combination.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen-plus-2025-07-28:thinking",
+    "name": "Qwen: Qwen Plus 0728 (thinking)",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-09-08 16:06:39 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.26,
+          "output_per_million": 0.78
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen Plus 0728, based on the Qwen3 foundation model, is a 1 million context hybrid reasoning model with a balanced performance, speed, and cost combination.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen-turbo",
+    "name": "Qwen: Qwen-Turbo",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-02-01 11:56:14 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.0325,
+          "output_per_million": 0.13,
+          "cache_read_input_per_million": 0.006500000000000001
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen-Turbo, based on Qwen2.5, is a 1M context model that provides fast speed and low cost, suitable for simple tasks.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 8192,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen-vl-max",
+    "name": "Qwen: Qwen VL Max",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-02-01 18:25:04 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.52,
+          "output_per_million": 2.08
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen VL Max is a visual understanding model with 7500 tokens context length. It excels in delivering optimal performance for a broader spectrum of complex tasks.\n",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen-vl-plus",
+    "name": "Qwen: Qwen VL Plus",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-02-05 04:54:15 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.1365,
+          "output_per_million": 0.40950000000000003,
+          "cache_read_input_per_million": 0.027299999999999998
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen's Enhanced Large Visual Language Model. Significantly upgraded for detailed recognition capabilities and text recognition abilities, supporting ultra-high pixel resolutions up to millions of pixels and extreme aspect ratios for...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 8192,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "temperature",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "qwen/qwen2.5-vl-72b-instruct",
     "name": "Qwen2.5 VL 72B Instruct",
     "provider": "openrouter",
     "family": "qwen",
-    "created_at": "2025-02-01 00:00:00 +0530",
+    "created_at": "2025-02-01 00:00:00 UTC",
     "context_window": 32768,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -20415,10 +50462,52 @@
     },
     "capabilities": [
       "structured_output",
-      "vision"
+      "vision",
+      "streaming",
+      "predicted_outputs"
     ],
-    "pricing": {},
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.25,
+          "output_per_million": 0.75
+        }
+      }
+    },
     "metadata": {
+      "description": "Qwen2.5-VL is proficient in recognizing common objects such as flowers, birds, fish, and insects. It is also highly capable of analyzing texts, charts, icons, graphics, and layouts within images.",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 32000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -20436,12 +50525,148 @@
       "knowledge": "2024-10"
     }
   },
+  {
+    "id": "qwen/qwen3-14b",
+    "name": "Qwen: Qwen3 14B",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-04-28 21:41:18 UTC",
+    "context_window": 40960,
+    "max_output_tokens": 40960,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.06,
+          "output_per_million": 0.24
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3-14B is a dense 14.8B parameter causal language model from the Qwen3 series, designed for both complex reasoning and efficient dialogue. It supports seamless switching between a \"thinking\" mode for...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": "qwen3"
+      },
+      "top_provider": {
+        "context_length": 40960,
+        "max_completion_tokens": 40960,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen3-235b-a22b",
+    "name": "Qwen: Qwen3 235B A22B",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-04-28 21:29:17 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.45499999999999996,
+          "output_per_million": 1.8199999999999998
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3-235B-A22B is a 235B parameter mixture-of-experts (MoE) model developed by Qwen, activating 22B parameters per forward pass. It supports seamless switching between a \"thinking\" mode for complex reasoning, math, and...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": "qwen3"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 8192,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "qwen/qwen3-235b-a22b-07-25",
     "name": "Qwen3 235B A22B Instruct 2507",
     "provider": "openrouter",
     "family": "qwen",
-    "created_at": "2025-04-28 00:00:00 +0530",
+    "created_at": "2025-04-28 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -20483,12 +50708,83 @@
       "knowledge": "2025-04"
     }
   },
+  {
+    "id": "qwen/qwen3-235b-a22b-2507",
+    "name": "Qwen: Qwen3 235B A22B Instruct 2507",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-07-21 17:39:15 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.071,
+          "output_per_million": 0.09999999999999999
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3-235B-A22B-Instruct-2507 is a multilingual, instruction-tuned mixture-of-experts language model based on the Qwen3-235B architecture, with 22B active parameters per forward pass. It is optimized for general-purpose text generation, including instruction following,...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "qwen/qwen3-235b-a22b-thinking-2507",
     "name": "Qwen3 235B A22B Thinking 2507",
     "provider": "openrouter",
     "family": "qwen",
-    "created_at": "2025-07-25 00:00:00 +0530",
+    "created_at": "2025-07-25 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 81920,
     "knowledge_cutoff": null,
@@ -20503,7 +50799,9 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -20514,6 +50812,43 @@
       }
     },
     "metadata": {
+      "description": "Qwen3-235B-A22B-Thinking-2507 is a high-performance, open-weight Mixture-of-Experts (MoE) language model optimized for complex reasoning tasks. It activates 22B of its 235B parameters per forward pass and natively supports up to 262,144...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": "qwen3"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -20531,12 +50866,85 @@
       "knowledge": "2025-04"
     }
   },
+  {
+    "id": "qwen/qwen3-30b-a3b",
+    "name": "Qwen: Qwen3 30B A3B",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-04-28 22:16:44 UTC",
+    "context_window": 40960,
+    "max_output_tokens": 20000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.09,
+          "output_per_million": 0.44999999999999996
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3, the latest generation in the Qwen large language model series, features both dense and mixture-of-experts (MoE) architectures to excel in reasoning, multilingual support, and advanced agent tasks. Its unique...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": "qwen3"
+      },
+      "top_provider": {
+        "context_length": 40960,
+        "max_completion_tokens": 20000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "qwen/qwen3-30b-a3b-instruct-2507",
     "name": "Qwen3 30B A3B Instruct 2507",
     "provider": "openrouter",
     "family": "qwen",
-    "created_at": "2025-07-29 00:00:00 +0530",
+    "created_at": "2025-07-29 00:00:00 UTC",
     "context_window": 262000,
     "max_output_tokens": 262000,
     "knowledge_cutoff": null,
@@ -20550,7 +50958,9 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -20561,6 +50971,41 @@
       }
     },
     "metadata": {
+      "description": "Qwen3-30B-A3B-Instruct-2507 is a 30.5B-parameter mixture-of-experts language model from Qwen, with 3.3B active parameters per inference. It operates in non-thinking mode and is designed for high-quality instruction following, multilingual understanding, and...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 262144,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -20583,7 +51028,7 @@
     "name": "Qwen3 30B A3B Thinking 2507",
     "provider": "openrouter",
     "family": "qwen",
-    "created_at": "2025-07-29 00:00:00 +0530",
+    "created_at": "2025-07-29 00:00:00 UTC",
     "context_window": 262000,
     "max_output_tokens": 262000,
     "knowledge_cutoff": null,
@@ -20598,7 +51043,9 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -20609,6 +51056,43 @@
       }
     },
     "metadata": {
+      "description": "Qwen3-30B-A3B-Thinking-2507 is a 30B parameter Mixture-of-Experts reasoning model optimized for complex tasks requiring extended multi-step thinking. The model is designed specifically for “thinking mode,” where internal reasoning traces are separated...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 131072,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -20626,12 +51110,156 @@
       "knowledge": "2025-04"
     }
   },
+  {
+    "id": "qwen/qwen3-32b",
+    "name": "Qwen: Qwen3 32B",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-04-28 21:32:25 UTC",
+    "context_window": 40960,
+    "max_output_tokens": 40960,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.08,
+          "output_per_million": 0.24,
+          "cache_read_input_per_million": 0.04
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3-32B is a dense 32.8B parameter causal language model from the Qwen3 series, optimized for both complex reasoning and efficient dialogue. It supports seamless switching between a \"thinking\" mode for...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": "qwen3"
+      },
+      "top_provider": {
+        "context_length": 40960,
+        "max_completion_tokens": 40960,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen3-8b",
+    "name": "Qwen: Qwen3 8B",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-04-28 21:43:52 UTC",
+    "context_window": 40960,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.049999999999999996,
+          "output_per_million": 0.39999999999999997,
+          "cache_read_input_per_million": 0.049999999999999996
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3-8B is a dense 8.2B parameter causal language model from the Qwen3 series, designed for both reasoning-heavy tasks and efficient dialogue. It supports seamless switching between \"thinking\" mode for math,...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": "qwen3"
+      },
+      "top_provider": {
+        "context_length": 40960,
+        "max_completion_tokens": 8192,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "qwen/qwen3-coder",
     "name": "Qwen3 Coder",
     "provider": "openrouter",
     "family": "qwen",
-    "created_at": "2025-07-23 00:00:00 +0530",
+    "created_at": "2025-07-23 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 66536,
     "knowledge_cutoff": null,
@@ -20645,7 +51273,9 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -20656,6 +51286,41 @@
       }
     },
     "metadata": {
+      "description": "Qwen3-Coder-480B-A35B-Instruct is a Mixture-of-Experts (MoE) code generation model developed by the Qwen team. It is optimized for agentic coding tasks such as function calling, tool use, and long-context reasoning over...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -20678,7 +51343,7 @@
     "name": "Qwen3 Coder 30B A3B Instruct",
     "provider": "openrouter",
     "family": "qwen",
-    "created_at": "2025-07-31 00:00:00 +0530",
+    "created_at": "2025-07-31 00:00:00 UTC",
     "context_window": 160000,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -20692,7 +51357,8 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -20703,6 +51369,39 @@
       }
     },
     "metadata": {
+      "description": "Qwen3-Coder-30B-A3B-Instruct is a 30.5B parameter Mixture-of-Experts (MoE) model with 128 experts (8 active per forward pass), designed for advanced code generation, repository-scale understanding, and agentic tool use. Built on the...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 160000,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -20725,7 +51424,7 @@
     "name": "Qwen3 Coder Flash",
     "provider": "openrouter",
     "family": "qwen",
-    "created_at": "2025-07-23 00:00:00 +0530",
+    "created_at": "2025-07-23 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 66536,
     "knowledge_cutoff": null,
@@ -20738,7 +51437,9 @@
       ]
     },
     "capabilities": [
-      "function_calling"
+      "function_calling",
+      "streaming",
+      "structured_output"
     ],
     "pricing": {
       "text_tokens": {
@@ -20749,6 +51450,34 @@
       }
     },
     "metadata": {
+      "description": "Qwen3 Coder Flash is Alibaba's fast and cost efficient version of their proprietary Qwen3 Coder Plus. It is a powerful coding agent model specializing in autonomous programming via tool calling...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -20766,12 +51495,145 @@
       "knowledge": "2025-04"
     }
   },
+  {
+    "id": "qwen/qwen3-coder-next",
+    "name": "Qwen: Qwen3 Coder Next",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2026-02-04 00:15:01 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 262144,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.11,
+          "output_per_million": 0.7999999999999999,
+          "cache_read_input_per_million": 0.07
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3-Coder-Next is an open-weight causal language model optimized for coding agents and local development workflows. It uses a sparse MoE design with 80B total parameters and only 3B activated per...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 262144,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen3-coder-plus",
+    "name": "Qwen: Qwen3 Coder Plus",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-09-23 21:25:07 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.65,
+          "output_per_million": 3.25,
+          "cache_read_input_per_million": 0.13
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3 Coder Plus is Alibaba's proprietary version of the Open Source Qwen3 Coder 480B A35B. It is a powerful coding agent model specializing in autonomous programming via tool calling and...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "qwen/qwen3-coder:exacto",
     "name": "Qwen3 Coder (exacto)",
     "provider": "openrouter",
     "family": "qwen",
-    "created_at": "2025-07-23 00:00:00 +0530",
+    "created_at": "2025-07-23 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -20813,12 +51675,66 @@
       "knowledge": "2025-04"
     }
   },
+  {
+    "id": "qwen/qwen3-coder:free",
+    "name": "Qwen: Qwen3 Coder 480B A35B (free)",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-07-23 00:29:06 UTC",
+    "context_window": 262000,
+    "max_output_tokens": 262000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "description": "Qwen3-Coder-480B-A35B-Instruct is a Mixture-of-Experts (MoE) code generation model developed by the Qwen team. It is optimized for agentic coding tasks such as function calling, tool use, and long-context reasoning over...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262000,
+        "max_completion_tokens": 262000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "qwen/qwen3-max",
     "name": "Qwen3 Max",
     "provider": "openrouter",
     "family": "qwen",
-    "created_at": "2025-09-05 00:00:00 +0530",
+    "created_at": "2025-09-05 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -20832,7 +51748,9 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "structured_output"
     ],
     "pricing": {
       "text_tokens": {
@@ -20843,6 +51761,34 @@
       }
     },
     "metadata": {
+      "description": "Qwen3-Max is an updated release built on the Qwen3 series, offering major improvements in reasoning, instruction following, multilingual support, and long-tail knowledge coverage compared to the January 2025 version. It...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -20860,13 +51806,13 @@
     }
   },
   {
-    "id": "qwen/qwen3-next-80b-a3b-instruct",
-    "name": "Qwen3 Next 80B A3B Instruct",
+    "id": "qwen/qwen3-max-thinking",
+    "name": "Qwen: Qwen3 Max Thinking",
     "provider": "openrouter",
     "family": "qwen",
-    "created_at": "2025-09-11 00:00:00 +0530",
+    "created_at": "2026-02-09 21:18:21 UTC",
     "context_window": 262144,
-    "max_output_tokens": 262144,
+    "max_output_tokens": 32768,
     "knowledge_cutoff": null,
     "modalities": {
       "input": [
@@ -20877,41 +51823,58 @@
       ]
     },
     "capabilities": [
+      "streaming",
       "function_calling",
       "structured_output"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
-          "input_per_million": 0.14,
-          "output_per_million": 1.4
+          "input_per_million": 0.78,
+          "output_per_million": 3.9
         }
       }
     },
     "metadata": {
-      "source": "models.dev",
-      "provider_id": "openrouter",
-      "open_weights": true,
-      "attachment": false,
-      "temperature": true,
-      "last_updated": "2025-09-11",
-      "cost": {
-        "input": 0.14,
-        "output": 1.4
+      "description": "Qwen3-Max-Thinking is the flagship reasoning model in the Qwen3 series, designed for high-stakes cognitive tasks that require deep, multi-step reasoning. By significantly scaling model capacity and reinforcement learning compute, it...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen",
+        "instruct_type": null
       },
-      "limit": {
-        "context": 262144,
-        "output": 262144
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
       },
-      "knowledge": "2025-04"
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
     }
   },
   {
-    "id": "qwen/qwen3-next-80b-a3b-thinking",
-    "name": "Qwen3 Next 80B A3B Thinking",
+    "id": "qwen/qwen3-next-80b-a3b-instruct",
+    "name": "Qwen3 Next 80B A3B Instruct",
     "provider": "openrouter",
     "family": "qwen",
-    "created_at": "2025-09-11 00:00:00 +0530",
+    "created_at": "2025-09-11 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 262144,
     "knowledge_cutoff": null,
@@ -20926,7 +51889,8 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -20937,6 +51901,41 @@
       }
     },
     "metadata": {
+      "description": "Qwen3-Next-80B-A3B-Instruct is an instruction-tuned chat model in the Qwen3-Next series optimized for fast, stable responses without “thinking” traces. It targets complex tasks across reasoning, code generation, knowledge QA, and multilingual...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -20954,12 +51953,870 @@
       "knowledge": "2025-04"
     }
   },
+  {
+    "id": "qwen/qwen3-next-80b-a3b-instruct:free",
+    "name": "Qwen: Qwen3 Next 80B A3B Instruct (free)",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-09-11 17:36:53 UTC",
+    "context_window": 262144,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {},
+    "metadata": {
+      "description": "Qwen3-Next-80B-A3B-Instruct is an instruction-tuned chat model in the Qwen3-Next series optimized for fast, stable responses without “thinking” traces. It targets complex tasks across reasoning, code generation, knowledge QA, and multilingual...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen3-next-80b-a3b-thinking",
+    "name": "Qwen3 Next 80B A3B Thinking",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-09-11 00:00:00 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 262144,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.14,
+          "output_per_million": 1.4
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3-Next-80B-A3B-Thinking is a reasoning-first chat model in the Qwen3-Next line that outputs structured “thinking” traces by default. It’s designed for hard multi-step problems; math proofs, code synthesis/debugging, logic, and agentic...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
+      "source": "models.dev",
+      "provider_id": "openrouter",
+      "open_weights": true,
+      "attachment": false,
+      "temperature": true,
+      "last_updated": "2025-09-11",
+      "cost": {
+        "input": 0.14,
+        "output": 1.4
+      },
+      "limit": {
+        "context": 262144,
+        "output": 262144
+      },
+      "knowledge": "2025-04"
+    }
+  },
+  {
+    "id": "qwen/qwen3-vl-235b-a22b-instruct",
+    "name": "Qwen: Qwen3 VL 235B A22B Instruct",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-09-23 23:04:47 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.19999999999999998,
+          "output_per_million": 0.88,
+          "cache_read_input_per_million": 0.11
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3-VL-235B-A22B Instruct is an open-weight multimodal model that unifies strong text generation with visual understanding across images and video. The Instruct model targets general vision-language use (VQA, document parsing, chart/table...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen3-vl-235b-a22b-thinking",
+    "name": "Qwen: Qwen3 VL 235B A22B Thinking",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-09-23 23:04:50 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.26,
+          "output_per_million": 2.6
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3-VL-235B-A22B Thinking is a multimodal model that unifies strong text generation with visual understanding across images and video. The Thinking model is optimized for multimodal reasoning in STEM and math....",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen3-vl-30b-a3b-instruct",
+    "name": "Qwen: Qwen3 VL 30B A3B Instruct",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-10-06 23:47:56 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.13,
+          "output_per_million": 0.52
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3-VL-30B-A3B-Instruct is a multimodal model that unifies strong text generation with visual understanding for images and videos. Its Instruct variant optimizes instruction-following for general multimodal tasks. It excels in perception...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen3-vl-30b-a3b-thinking",
+    "name": "Qwen: Qwen3 VL 30B A3B Thinking",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-10-06 23:47:59 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.13,
+          "output_per_million": 1.56
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3-VL-30B-A3B-Thinking is a multimodal model that unifies strong text generation with visual understanding for images and videos. Its Thinking variant enhances reasoning in STEM, math, and complex tasks. It excels...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen3-vl-32b-instruct",
+    "name": "Qwen: Qwen3 VL 32B Instruct",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-10-23 14:55:32 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.10400000000000001,
+          "output_per_million": 0.41600000000000004
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3-VL-32B-Instruct is a large-scale multimodal vision-language model designed for high-precision understanding and reasoning across text, images, and video. With 32 billion parameters, it combines deep visual perception with advanced text...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen3-vl-8b-instruct",
+    "name": "Qwen: Qwen3 VL 8B Instruct",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-10-14 17:35:08 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.08,
+          "output_per_million": 0.5
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3-VL-8B-Instruct is a multimodal vision-language model from the Qwen3-VL series, built for high-fidelity understanding and reasoning across text, images, and video. It features improved multimodal fusion with Interleaved-MRoPE for long-horizon...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen3-vl-8b-thinking",
+    "name": "Qwen: Qwen3 VL 8B Thinking",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2025-10-14 17:42:26 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.117,
+          "output_per_million": 1.365
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3-VL-8B-Thinking is the reasoning-optimized variant of the Qwen3-VL-8B multimodal model, designed for advanced visual and textual reasoning across complex scenes, documents, and temporal sequences. It integrates enhanced multimodal alignment and...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen3.5-122b-a10b",
+    "name": "Qwen: Qwen3.5-122B-A10B",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2026-02-25 21:09:49 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.26,
+          "output_per_million": 2.08
+        }
+      }
+    },
+    "metadata": {
+      "description": "The Qwen3.5 122B-A10B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. In terms of...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen3.5-27b",
+    "name": "Qwen: Qwen3.5-27B",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2026-02-25 21:10:10 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.195,
+          "output_per_million": 1.56
+        }
+      }
+    },
+    "metadata": {
+      "description": "The Qwen3.5 27B native vision-language Dense model incorporates a linear attention mechanism, delivering fast response times while balancing inference speed and performance. Its overall capabilities are comparable to those of...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen3.5-35b-a3b",
+    "name": "Qwen: Qwen3.5-35B-A3B",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2026-02-25 21:10:22 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 262144,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.15,
+          "output_per_million": 1.0,
+          "cache_read_input_per_million": 0.049999999999999996
+        }
+      }
+    },
+    "metadata": {
+      "description": "The Qwen3.5 Series 35B-A3B is a native vision-language model designed with a hybrid architecture that integrates linear attention mechanisms and a sparse mixture-of-experts model, achieving higher inference efficiency. Its overall...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 262144,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "qwen/qwen3.5-397b-a17b",
     "name": "Qwen3.5 397B A17B",
     "provider": "openrouter",
     "family": "qwen",
-    "created_at": "2026-02-16 00:00:00 +0530",
+    "created_at": "2026-02-16 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -20977,7 +52834,9 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -20988,6 +52847,47 @@
       }
     },
     "metadata": {
+      "description": "The Qwen3.5 series 397B-A17B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. It delivers...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -21005,12 +52905,88 @@
       "knowledge": "2025-04"
     }
   },
+  {
+    "id": "qwen/qwen3.5-9b",
+    "name": "Qwen: Qwen3.5-9B",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2026-03-10 14:19:56 UTC",
+    "context_window": 262144,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.09999999999999999,
+          "output_per_million": 0.15
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3.5-9B is a multimodal foundation model from the Qwen3.5 family, designed to deliver strong reasoning, coding, and visual understanding in an efficient 9B-parameter architecture. It uses a unified vision-language design...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "qwen/qwen3.5-flash-02-23",
     "name": "Qwen: Qwen3.5-Flash",
     "provider": "openrouter",
     "family": "qwen",
-    "created_at": "2026-02-25 00:00:00 +0530",
+    "created_at": "2026-02-25 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -21028,7 +53004,8 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -21039,6 +53016,39 @@
       }
     },
     "metadata": {
+      "description": "The Qwen3.5 native vision-language Flash models are built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. Compared to the...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -21060,7 +53070,7 @@
     "name": "Qwen3.5 Plus 2026-02-15",
     "provider": "openrouter",
     "family": "qwen",
-    "created_at": "2026-02-16 00:00:00 +0530",
+    "created_at": "2026-02-16 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -21078,7 +53088,8 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -21089,6 +53100,39 @@
       }
     },
     "metadata": {
+      "description": "The Qwen3.5 native vision-language series Plus models are built on a hybrid architecture that integrates linear attention mechanisms with sparse mixture-of-experts models, achieving higher inference efficiency. In a variety of...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -21106,12 +53150,367 @@
       "knowledge": "2025-04"
     }
   },
+  {
+    "id": "qwen/qwen3.5-plus-20260420",
+    "name": "Qwen: Qwen3.5 Plus 2026-04-20",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2026-04-27 03:42:48 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.39999999999999997,
+          "output_per_million": 2.4
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3.5 Plus (April 2026) is a large-scale multimodal language model from Alibaba. It accepts text, image, and video input and produces text output, with a 1M token context window. This...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen3.6-27b",
+    "name": "Qwen: Qwen3.6 27B",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2026-04-27 01:57:44 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 81920,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.32,
+          "output_per_million": 3.1999999999999997
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3.6 27B is a dense 27-billion-parameter language model from the Qwen Team at Alibaba, released in April 2026. It features hybrid multimodal capabilities — accepting text, image, and video inputs...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 81920,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen3.6-35b-a3b",
+    "name": "Qwen: Qwen3.6 35B A3B",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2026-04-27 03:24:15 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 262144,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.15,
+          "output_per_million": 1.0,
+          "cache_read_input_per_million": 0.049999999999999996
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3.6-35B-A3B is an open-weight multimodal model from Alibaba Cloud with 35 billion total parameters and 3 billion active parameters per token. It uses a hybrid sparse mixture-of-experts architecture combining Gated...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 262144,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen3.6-flash",
+    "name": "Qwen: Qwen3.6 Flash",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2026-04-27 03:42:42 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.25,
+          "output_per_million": 1.5
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3.6 Flash is a fast, efficient language model from Alibaba's Qwen 3.6 series. It supports text, image, and video input with a 1M token context window. Tiered pricing kicks in...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "qwen/qwen3.6-max-preview",
+    "name": "Qwen: Qwen3.6 Max Preview",
+    "provider": "openrouter",
+    "family": "qwen",
+    "created_at": "2026-04-27 03:24:02 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.04,
+          "output_per_million": 6.24
+        }
+      }
+    },
+    "metadata": {
+      "description": "Qwen3.6-Max-Preview is a proprietary frontier model from Alibaba Cloud built on a sparse mixture-of-experts architecture with approximately 1 trillion total parameters. It is optimized for agentic coding, tool use, and...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "qwen/qwen3.6-plus",
     "name": "Qwen3.6 Plus",
     "provider": "openrouter",
     "family": "qwen",
-    "created_at": "2026-04-02 00:00:00 +0530",
+    "created_at": "2026-04-02 00:00:00 UTC",
     "context_window": 1000000,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -21129,7 +53528,8 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -21140,6 +53540,39 @@
       }
     },
     "metadata": {
+      "description": "Qwen 3.6 Plus builds on a hybrid architecture that combines efficient linear attention with sparse mixture-of-experts routing, enabling strong scalability and high-performance inference. Compared to the 3.5 series, it delivers...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -21157,12 +53590,582 @@
       "knowledge": "2025-04"
     }
   },
+  {
+    "id": "rekaai/reka-edge",
+    "name": "Reka Edge",
+    "provider": "openrouter",
+    "family": "rekaai",
+    "created_at": "2026-03-20 17:16:05 UTC",
+    "context_window": 16384,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.09999999999999999,
+          "output_per_million": 0.09999999999999999
+        }
+      }
+    },
+    "metadata": {
+      "description": "Reka Edge is an extremely efficient 7B multimodal vision-language model that accepts image/video+text inputs and generates text outputs. This model is optimized specifically to deliver industry-leading performance in image understanding,...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 16384,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "rekaai/reka-flash-3",
+    "name": "Reka Flash 3",
+    "provider": "openrouter",
+    "family": "rekaai",
+    "created_at": "2025-03-12 20:53:33 UTC",
+    "context_window": 65536,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.09999999999999999,
+          "output_per_million": 0.19999999999999998
+        }
+      }
+    },
+    "metadata": {
+      "description": "Reka Flash 3 is a general-purpose, instruction-tuned large language model with 21 billion parameters, developed by Reka. It excels at general chat, coding tasks, instruction-following, and function calling. Featuring a...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 65536,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "relace/relace-apply-3",
+    "name": "Relace: Relace Apply 3",
+    "provider": "openrouter",
+    "family": "relace",
+    "created_at": "2025-09-26 12:59:32 UTC",
+    "context_window": 256000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.85,
+          "output_per_million": 1.25
+        }
+      }
+    },
+    "metadata": {
+      "description": "Relace Apply 3 is a specialized code-patching LLM that merges AI-suggested edits straight into your source files. It can apply updates from GPT-4o, Claude, and others into your files at...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 256000,
+        "max_completion_tokens": 128000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "seed",
+        "stop"
+      ]
+    }
+  },
+  {
+    "id": "relace/relace-search",
+    "name": "Relace: Relace Search",
+    "provider": "openrouter",
+    "family": "relace",
+    "created_at": "2025-12-08 17:06:00 UTC",
+    "context_window": 256000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.0,
+          "output_per_million": 3.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "The relace-search model uses 4-12 `view_file` and `grep` tools in parallel to explore a codebase and return relevant files to the user request. In contrast to RAG, relace-search performs agentic...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 256000,
+        "max_completion_tokens": 128000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "sao10k/l3-euryale-70b",
+    "name": "Sao10k: Llama 3 Euryale 70B v2.1",
+    "provider": "openrouter",
+    "family": "sao10k",
+    "created_at": "2024-06-18 00:00:00 UTC",
+    "context_window": 8192,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.48,
+          "output_per_million": 1.48
+        }
+      }
+    },
+    "metadata": {
+      "description": "Euryale 70B v2.1 is a model focused on creative roleplay from [Sao10k](https://ko-fi.com/sao10k). - Better prompt adherence. - Better anatomy / spatial awareness. - Adapts much better to unique and custom...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "llama3"
+      },
+      "top_provider": {
+        "context_length": 8192,
+        "max_completion_tokens": 8192,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "sao10k/l3-lunaris-8b",
+    "name": "Sao10K: Llama 3 8B Lunaris",
+    "provider": "openrouter",
+    "family": "sao10k",
+    "created_at": "2024-08-13 00:00:00 UTC",
+    "context_window": 8192,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.04,
+          "output_per_million": 0.049999999999999996
+        }
+      }
+    },
+    "metadata": {
+      "description": "Lunaris 8B is a versatile generalist and roleplaying model based on Llama 3. It's a strategic merge of multiple models, designed to balance creativity with improved logic and general knowledge....",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "llama3"
+      },
+      "top_provider": {
+        "context_length": 8192,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "sao10k/l3.1-70b-hanami-x1",
+    "name": "Sao10K: Llama 3.1 70B Hanami x1",
+    "provider": "openrouter",
+    "family": "sao10k",
+    "created_at": "2025-01-08 02:20:54 UTC",
+    "context_window": 16000,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 3.0,
+          "output_per_million": 3.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "This is [Sao10K](/sao10k)'s experiment over [Euryale v2.2](/sao10k/l3.1-euryale-70b).",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 16000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "sao10k/l3.1-euryale-70b",
+    "name": "Sao10K: Llama 3.1 Euryale 70B v2.2",
+    "provider": "openrouter",
+    "family": "sao10k",
+    "created_at": "2024-08-28 00:00:00 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.85,
+          "output_per_million": 0.85
+        }
+      }
+    },
+    "metadata": {
+      "description": "Euryale L3.1 70B v2.2 is a model focused on creative roleplay from [Sao10k](https://ko-fi.com/sao10k). It is the successor of [Euryale L3 70B v2.1](/models/sao10k/l3-euryale-70b).",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "llama3"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "sao10k/l3.3-euryale-70b",
+    "name": "Sao10K: Llama 3.3 Euryale 70B",
+    "provider": "openrouter",
+    "family": "sao10k",
+    "created_at": "2024-12-18 15:32:08 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.65,
+          "output_per_million": 0.75
+        }
+      }
+    },
+    "metadata": {
+      "description": "Euryale L3.3 70B is a model focused on creative roleplay from [Sao10k](https://ko-fi.com/sao10k). It is the successor of [Euryale L3 70B v2.2](/models/sao10k/l3-euryale-70b).",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama3",
+        "instruct_type": "llama3"
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "sourceful/riverflow-v2-fast-preview",
     "name": "Riverflow V2 Fast Preview",
     "provider": "openrouter",
     "family": "sourceful",
-    "created_at": "2025-12-08 00:00:00 +0530",
+    "created_at": "2025-12-08 00:00:00 UTC",
     "context_window": 8192,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -21202,7 +54205,7 @@
     "name": "Riverflow V2 Max Preview",
     "provider": "openrouter",
     "family": "sourceful",
-    "created_at": "2025-12-08 00:00:00 +0530",
+    "created_at": "2025-12-08 00:00:00 UTC",
     "context_window": 8192,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -21242,7 +54245,7 @@
     "name": "Riverflow V2 Standard Preview",
     "provider": "openrouter",
     "family": "sourceful",
-    "created_at": "2025-12-08 00:00:00 +0530",
+    "created_at": "2025-12-08 00:00:00 UTC",
     "context_window": 8192,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -21282,7 +54285,7 @@
     "name": "Step 3.5 Flash",
     "provider": "openrouter",
     "family": "step",
-    "created_at": "2026-01-29 00:00:00 +0530",
+    "created_at": "2026-01-29 00:00:00 UTC",
     "context_window": 256000,
     "max_output_tokens": 256000,
     "knowledge_cutoff": null,
@@ -21296,18 +54299,57 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.1,
           "output_per_million": 0.3,
-          "cached_input_per_million": 0.02
+          "cache_read_input_per_million": 0.02
         }
       }
     },
     "metadata": {
+      "description": "Step 3.5 Flash is StepFun's most capable open-source foundation model. Built on a sparse Mixture of Experts (MoE) architecture, it selectively activates only 11B of its 196B parameters per token....",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -21326,12 +54368,708 @@
       "knowledge": "2025-01"
     }
   },
+  {
+    "id": "switchpoint/router",
+    "name": "Switchpoint Router",
+    "provider": "openrouter",
+    "family": "switchpoint",
+    "created_at": "2025-07-11 22:28:19 UTC",
+    "context_window": 131072,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.85,
+          "output_per_million": 3.4
+        }
+      }
+    },
+    "metadata": {
+      "description": "Switchpoint AI's router instantly analyzes your request and directs it to the optimal AI from an ever-evolving library. As the world of LLMs advances, our router gets smarter, ensuring you...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "tencent/hunyuan-a13b-instruct",
+    "name": "Tencent: Hunyuan A13B Instruct",
+    "provider": "openrouter",
+    "family": "tencent",
+    "created_at": "2025-07-08 15:14:24 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 131072,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.14,
+          "output_per_million": 0.5700000000000001
+        }
+      }
+    },
+    "metadata": {
+      "description": "Hunyuan-A13B is a 13B active parameter Mixture-of-Experts (MoE) language model developed by Tencent, with a total parameter count of 80B and support for reasoning via Chain-of-Thought. It offers competitive benchmark...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 131072,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "structured_outputs",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "tencent/hy3-preview:free",
+    "name": "Tencent: Hy3 preview (free)",
+    "provider": "openrouter",
+    "family": "tencent",
+    "created_at": "2026-04-22 17:15:50 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 262144,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "description": "Hy3 preview is a high-efficiency Mixture-of-Experts model from Tencent designed for agentic workflows and production use. It supports configurable reasoning levels across disabled, low, and high modes, allowing it to...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 262144,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "thedrummer/cydonia-24b-v4.1",
+    "name": "TheDrummer: Cydonia 24B V4.1",
+    "provider": "openrouter",
+    "family": "thedrummer",
+    "created_at": "2025-09-27 00:11:18 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 131072,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.3,
+          "output_per_million": 0.5,
+          "cache_read_input_per_million": 0.15
+        }
+      }
+    },
+    "metadata": {
+      "description": "Uncensored and creative writing model based on Mistral Small 3.2 24B with good recall, prompt adherence, and intelligence.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 131072,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "thedrummer/rocinante-12b",
+    "name": "TheDrummer: Rocinante 12B",
+    "provider": "openrouter",
+    "family": "thedrummer",
+    "created_at": "2024-09-30 00:00:00 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.16999999999999998,
+          "output_per_million": 0.43
+        }
+      }
+    },
+    "metadata": {
+      "description": "Rocinante 12B is designed for engaging storytelling and rich prose. Early testers have reported: - Expanded vocabulary with unique and expressive word choices - Enhanced creativity for vivid narratives -...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Qwen",
+        "instruct_type": "chatml"
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "thedrummer/skyfall-36b-v2",
+    "name": "TheDrummer: Skyfall 36B V2",
+    "provider": "openrouter",
+    "family": "thedrummer",
+    "created_at": "2025-03-10 19:56:06 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.55,
+          "output_per_million": 0.7999999999999999,
+          "cache_read_input_per_million": 0.25
+        }
+      }
+    },
+    "metadata": {
+      "description": "Skyfall 36B v2 is an enhanced iteration of Mistral Small 2501, specifically fine-tuned for improved creativity, nuanced writing, role-playing, and coherent storytelling.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "thedrummer/unslopnemo-12b",
+    "name": "TheDrummer: UnslopNemo 12B",
+    "provider": "openrouter",
+    "family": "thedrummer",
+    "created_at": "2024-11-08 22:04:08 UTC",
+    "context_window": 32768,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.39999999999999997,
+          "output_per_million": 0.39999999999999997
+        }
+      }
+    },
+    "metadata": {
+      "description": "UnslopNemo v4.1 is the latest addition from the creator of Rocinante, designed for adventure writing and role-play scenarios.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Mistral",
+        "instruct_type": "mistral"
+      },
+      "top_provider": {
+        "context_length": 32768,
+        "max_completion_tokens": 32768,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "tngtech/deepseek-r1t2-chimera",
+    "name": "TNG: DeepSeek R1T2 Chimera",
+    "provider": "openrouter",
+    "family": "tngtech",
+    "created_at": "2025-07-08 15:03:05 UTC",
+    "context_window": 163840,
+    "max_output_tokens": 163840,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.3,
+          "output_per_million": 1.1,
+          "cache_read_input_per_million": 0.15
+        }
+      }
+    },
+    "metadata": {
+      "description": "DeepSeek-TNG-R1T2-Chimera is the second-generation Chimera model from TNG Tech. It is a 671 B-parameter mixture-of-experts text-generation model assembled from DeepSeek-AI’s R1-0528, R1, and V3-0324 checkpoints with an Assembly-of-Experts merge. The...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "DeepSeek",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 163840,
+        "max_completion_tokens": 163840,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "undi95/remm-slerp-l2-13b",
+    "name": "ReMM SLERP 13B",
+    "provider": "openrouter",
+    "family": "undi95",
+    "created_at": "2023-07-22 00:00:00 UTC",
+    "context_window": 6144,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.44999999999999996,
+          "output_per_million": 0.65
+        }
+      }
+    },
+    "metadata": {
+      "description": "A recreation trial of the original MythoMax-L2-B13 but with updated models. #merge",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Llama2",
+        "instruct_type": "alpaca"
+      },
+      "top_provider": {
+        "context_length": 6144,
+        "max_completion_tokens": 4096,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "top_a",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "upstage/solar-pro-3",
+    "name": "Upstage: Solar Pro 3",
+    "provider": "openrouter",
+    "family": "upstage",
+    "created_at": "2026-01-27 02:33:20 UTC",
+    "context_window": 128000,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.15,
+          "output_per_million": 0.6,
+          "cache_read_input_per_million": 0.015
+        }
+      }
+    },
+    "metadata": {
+      "description": "Solar Pro 3 is Upstage's powerful Mixture-of-Experts (MoE) language model. With 102B total parameters and 12B active parameters per forward pass, it delivers exceptional performance while maintaining computational efficiency. Optimized...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools"
+      ]
+    }
+  },
+  {
+    "id": "writer/palmyra-x5",
+    "name": "Writer: Palmyra X5",
+    "provider": "openrouter",
+    "family": "writer",
+    "created_at": "2026-01-21 13:57:03 UTC",
+    "context_window": 1040000,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.6,
+          "output_per_million": 6.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "Palmyra X5 is Writer's most advanced model, purpose-built for building and scaling AI agents across the enterprise. It delivers industry-leading speed and efficiency on context windows up to 1 million...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1040000,
+        "max_completion_tokens": 8192,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "stop",
+        "temperature",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "x-ai/grok-3",
     "name": "Grok 3",
     "provider": "openrouter",
     "family": "grok",
-    "created_at": "2025-02-17 00:00:00 +0530",
+    "created_at": "2025-02-17 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -21345,18 +55083,53 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output"
+      "structured_output",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.75
+          "cache_read_input_per_million": 0.75,
+          "cache_write_input_per_million": 15
         }
       }
     },
     "metadata": {
+      "description": "Grok 3 is the latest model from xAI. It's their flagship model that excels at enterprise use cases like data extraction, coding, and text summarization. Possesses deep domain knowledge in...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Grok",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -21381,7 +55154,7 @@
     "name": "Grok 3 Beta",
     "provider": "openrouter",
     "family": "grok",
-    "created_at": "2025-02-17 00:00:00 +0530",
+    "created_at": "2025-02-17 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -21394,18 +55167,53 @@
       ]
     },
     "capabilities": [
-      "function_calling"
+      "function_calling",
+      "streaming",
+      "structured_output"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.75
+          "cache_read_input_per_million": 0.75,
+          "cache_write_input_per_million": 15
         }
       }
     },
     "metadata": {
+      "description": "Grok 3 is the latest model from xAI. It's their flagship model that excels at enterprise use cases like data extraction, coding, and text summarization. Possesses deep domain knowledge in...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Grok",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -21430,7 +55238,7 @@
     "name": "Grok 3 Mini",
     "provider": "openrouter",
     "family": "grok",
-    "created_at": "2025-02-17 00:00:00 +0530",
+    "created_at": "2025-02-17 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -21445,18 +55253,53 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.3,
           "output_per_million": 0.5,
-          "cached_input_per_million": 0.075
+          "cache_read_input_per_million": 0.075,
+          "cache_write_input_per_million": 0.5
         }
       }
     },
     "metadata": {
+      "description": "A lightweight model that thinks before responding. Fast, smart, and great for logic-based tasks that do not require deep domain knowledge. The raw thinking traces are accessible.",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Grok",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "logprobs",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -21481,7 +55324,7 @@
     "name": "Grok 3 Mini Beta",
     "provider": "openrouter",
     "family": "grok",
-    "created_at": "2025-02-17 00:00:00 +0530",
+    "created_at": "2025-02-17 00:00:00 UTC",
     "context_window": 131072,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -21495,18 +55338,53 @@
     },
     "capabilities": [
       "function_calling",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "structured_output"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.3,
           "output_per_million": 0.5,
-          "cached_input_per_million": 0.075
+          "cache_read_input_per_million": 0.075,
+          "cache_write_input_per_million": 0.5
         }
       }
     },
     "metadata": {
+      "description": "Grok 3 Mini is a lightweight, smaller thinking model. Unlike traditional models that generate answers immediately, Grok 3 Mini thinks before responding. It’s ideal for reasoning-heavy tasks that don’t demand...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Grok",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "logprobs",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -21531,7 +55409,7 @@
     "name": "Grok 4",
     "provider": "openrouter",
     "family": "grok",
-    "created_at": "2025-07-09 00:00:00 +0530",
+    "created_at": "2025-07-09 00:00:00 UTC",
     "context_window": 256000,
     "max_output_tokens": 64000,
     "knowledge_cutoff": null,
@@ -21546,18 +55424,54 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 3,
           "output_per_million": 15,
-          "cached_input_per_million": 0.75
+          "cache_read_input_per_million": 0.75,
+          "cache_write_input_per_million": 15
         }
       }
     },
     "metadata": {
+      "description": "Grok 4 is xAI's latest reasoning model with a 256k context window. It supports parallel tool calling, structured outputs, and both image and text inputs. Note that reasoning is not...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Grok",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 256000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "logprobs",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -21582,7 +55496,7 @@
     "name": "Grok 4 Fast",
     "provider": "openrouter",
     "family": "grok",
-    "created_at": "2025-08-19 00:00:00 +0530",
+    "created_at": "2025-08-19 00:00:00 UTC",
     "context_window": 2000000,
     "max_output_tokens": 30000,
     "knowledge_cutoff": null,
@@ -21599,18 +55513,54 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.2,
           "output_per_million": 0.5,
-          "cached_input_per_million": 0.05
+          "cache_read_input_per_million": 0.05,
+          "cache_write_input_per_million": 0.05
         }
       }
     },
     "metadata": {
+      "description": "Grok 4 Fast is xAI's latest multimodal model with SOTA cost-efficiency and a 2M token context window. It comes in two flavors: non-reasoning and reasoning. Read more about the model...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Grok",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 2000000,
+        "max_completion_tokens": 30000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "logprobs",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -21635,7 +55585,7 @@
     "name": "Grok 4.1 Fast",
     "provider": "openrouter",
     "family": "grok",
-    "created_at": "2025-11-19 00:00:00 +0530",
+    "created_at": "2025-11-19 00:00:00 UTC",
     "context_window": 2000000,
     "max_output_tokens": 30000,
     "knowledge_cutoff": null,
@@ -21652,18 +55602,54 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.2,
           "output_per_million": 0.5,
-          "cached_input_per_million": 0.05
+          "cache_read_input_per_million": 0.05,
+          "cache_write_input_per_million": 0.05
         }
       }
     },
     "metadata": {
+      "description": "Grok 4.1 Fast is xAI's best agentic tool calling model that shines in real-world use cases like customer support and deep research. 2M context window. Reasoning can be enabled/disabled using...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Grok",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 2000000,
+        "max_completion_tokens": 30000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "logprobs",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -21683,12 +55669,82 @@
       "knowledge": "2024-11"
     }
   },
+  {
+    "id": "x-ai/grok-4.20",
+    "name": "xAI: Grok 4.20",
+    "provider": "openrouter",
+    "family": "x-ai",
+    "created_at": "2026-03-31 17:43:39 UTC",
+    "context_window": 2000000,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 2.5,
+          "cache_read_input_per_million": 0.19999999999999998
+        }
+      }
+    },
+    "metadata": {
+      "description": "Grok 4.20 is xAI's newest flagship model with industry-leading speed and agentic tool calling capabilities. It combines the lowest hallucination rate on the market with strict prompt adherance, delivering consistently...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Grok",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 2000000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "logprobs",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "x-ai/grok-4.20-beta",
     "name": "Grok 4.20 Beta",
     "provider": "openrouter",
     "family": "grok",
-    "created_at": "2026-03-12 00:00:00 +0530",
+    "created_at": "2026-03-12 00:00:00 UTC",
     "context_window": 2000000,
     "max_output_tokens": 30000,
     "knowledge_cutoff": null,
@@ -21711,7 +55767,7 @@
         "standard": {
           "input_per_million": 2,
           "output_per_million": 6,
-          "cached_input_per_million": 0.2
+          "cache_read_input_per_million": 0.2
         }
       }
     },
@@ -21738,12 +55794,79 @@
       }
     }
   },
+  {
+    "id": "x-ai/grok-4.20-multi-agent",
+    "name": "xAI: Grok 4.20 Multi-Agent",
+    "provider": "openrouter",
+    "family": "x-ai",
+    "created_at": "2026-03-31 17:45:58 UTC",
+    "context_window": 2000000,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "file"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "output_per_million": 6.0,
+          "cache_read_input_per_million": 0.19999999999999998
+        }
+      }
+    },
+    "metadata": {
+      "description": "Grok 4.20 Multi-Agent is a variant of xAI’s Grok 4.20 designed for collaborative, agent-based workflows. Multiple agents operate in parallel to conduct deep research, coordinate tool use, and synthesize information...",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Grok",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 2000000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "logprobs",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "temperature",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "x-ai/grok-4.20-multi-agent-beta",
     "name": "Grok 4.20 Multi - Agent Beta",
     "provider": "openrouter",
     "family": "grok",
-    "created_at": "2026-03-12 00:00:00 +0530",
+    "created_at": "2026-03-12 00:00:00 UTC",
     "context_window": 2000000,
     "max_output_tokens": 30000,
     "knowledge_cutoff": null,
@@ -21765,7 +55888,7 @@
         "standard": {
           "input_per_million": 2,
           "output_per_million": 6,
-          "cached_input_per_million": 0.2
+          "cache_read_input_per_million": 0.2
         }
       }
     },
@@ -21792,12 +55915,105 @@
       }
     }
   },
+  {
+    "id": "x-ai/grok-4.3",
+    "name": "Grok 4.3",
+    "provider": "openrouter",
+    "family": "grok",
+    "created_at": "2026-05-01 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 1000000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 2.5,
+          "cache_read_input_per_million": 0.2
+        }
+      }
+    },
+    "metadata": {
+      "description": "Grok 4.3 is a reasoning model from xAI. It accepts text and image inputs with text output, and is suited for agentic workflows, instruction-following tasks, and applications requiring high factual...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Grok",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logprobs",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ],
+      "source": "models.dev",
+      "provider_id": "openrouter",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2026-05-01",
+      "cost": {
+        "input": 1.25,
+        "output": 2.5,
+        "cache_read": 0.2,
+        "context_over_200k": {
+          "input": 2.5,
+          "output": 5,
+          "cache_read": 0.4
+        }
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 1000000
+      }
+    }
+  },
   {
     "id": "x-ai/grok-code-fast-1",
     "name": "Grok Code Fast 1",
     "provider": "openrouter",
     "family": "grok",
-    "created_at": "2025-08-26 00:00:00 +0530",
+    "created_at": "2025-08-26 00:00:00 UTC",
     "context_window": 256000,
     "max_output_tokens": 10000,
     "knowledge_cutoff": null,
@@ -21812,18 +56028,52 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.2,
           "output_per_million": 1.5,
-          "cached_input_per_million": 0.02
+          "cache_read_input_per_million": 0.02
         }
       }
     },
     "metadata": {
+      "description": "Grok Code Fast 1 is a speedy and economical reasoning model that excels at agentic coding. With reasoning traces visible in the response, developers can steer Grok Code for high-quality...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Grok",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 256000,
+        "max_completion_tokens": 10000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "logprobs",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -21844,13 +56094,13 @@
   },
   {
     "id": "xiaomi/mimo-v2-flash",
-    "name": "MiMo-V2-Flash",
+    "name": "Xiaomi: MiMo-V2-Flash",
     "provider": "openrouter",
     "family": "mimo",
-    "created_at": "2025-12-14 00:00:00 +0530",
+    "created_at": "2025-12-16 00:00:00 UTC",
     "context_window": 262144,
     "max_output_tokens": 65536,
-    "knowledge_cutoff": null,
+    "knowledge_cutoff": "2024-12-01",
     "modalities": {
       "input": [
         "text"
@@ -21861,25 +56111,64 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "structured_output"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.1,
           "output_per_million": 0.3,
-          "cached_input_per_million": 0.01
+          "cache_read_input_per_million": 0.01
         }
       }
     },
     "metadata": {
+      "description": "MiMo-V2-Flash is an open-source foundation language model developed by Xiaomi. It is a Mixture-of-Experts model with 309B total parameters and 15B active parameters, adopting hybrid attention architecture. MiMo-V2-Flash supports a...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
       "attachment": false,
       "temperature": true,
-      "last_updated": "2025-12-14",
+      "last_updated": "2026-02-04",
+      "interleaved": {
+        "field": "reasoning_details"
+      },
       "cost": {
         "input": 0.1,
         "output": 0.3,
@@ -21889,24 +56178,25 @@
         "context": 262144,
         "output": 65536
       },
-      "knowledge": "2024-12"
+      "knowledge": "2024-12-01"
     }
   },
   {
     "id": "xiaomi/mimo-v2-omni",
-    "name": "MiMo-V2-Omni",
+    "name": "Xiaomi: MiMo-V2-Omni",
     "provider": "openrouter",
     "family": "mimo",
-    "created_at": "2026-03-18 00:00:00 +0530",
+    "created_at": "2026-03-18 00:00:00 UTC",
     "context_window": 262144,
-    "max_output_tokens": 65536,
+    "max_output_tokens": 131072,
     "knowledge_cutoff": null,
     "modalities": {
       "input": [
         "text",
         "image",
+        "audio",
         "video",
-        "audio"
+        "pdf"
       ],
       "output": [
         "text"
@@ -21914,23 +56204,58 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming",
+      "structured_output"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.4,
           "output_per_million": 2,
-          "cached_input_per_million": 0.08
+          "cache_read_input_per_million": 0.08
         }
       }
     },
     "metadata": {
+      "description": "MiMo-V2-Omni is a frontier omni-modal model that natively processes image, video, and audio inputs within a unified architecture. It combines strong multimodal perception with agentic capability - visual grounding, multi-step...",
+      "architecture": {
+        "modality": "text+image+audio+video->text",
+        "input_modalities": [
+          "text",
+          "audio",
+          "image",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
-      "open_weights": true,
+      "open_weights": false,
       "attachment": true,
       "temperature": true,
       "last_updated": "2026-03-18",
@@ -21944,18 +56269,19 @@
       },
       "limit": {
         "context": 262144,
-        "output": 65536
-      }
+        "output": 131072
+      },
+      "knowledge": "2024-12"
     }
   },
   {
     "id": "xiaomi/mimo-v2-pro",
-    "name": "MiMo-V2-Pro",
+    "name": "Xiaomi: MiMo-V2-Pro",
     "provider": "openrouter",
     "family": "mimo",
-    "created_at": "2026-03-18 00:00:00 +0530",
+    "created_at": "2026-03-18 00:00:00 UTC",
     "context_window": 1048576,
-    "max_output_tokens": 65536,
+    "max_output_tokens": 131072,
     "knowledge_cutoff": null,
     "modalities": {
       "input": [
@@ -21967,23 +56293,55 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "structured_output"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1,
           "output_per_million": 3,
-          "cached_input_per_million": 0.2
+          "cache_read_input_per_million": 0.2
         }
       }
     },
     "metadata": {
+      "description": "MiMo-V2-Pro is Xiaomi's flagship foundation model, featuring over 1T total parameters and a 1M context length, deeply optimized for agentic scenarios. It is highly adaptable to general agent frameworks like...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 131072,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
-      "open_weights": true,
-      "attachment": false,
+      "open_weights": false,
+      "attachment": true,
       "temperature": true,
       "last_updated": "2026-03-18",
       "interleaved": {
@@ -21992,12 +56350,262 @@
       "cost": {
         "input": 1,
         "output": 3,
-        "cache_read": 0.2
+        "cache_read": 0.2,
+        "context_over_200k": {
+          "input": 2,
+          "output": 6,
+          "cache_read": 0.4
+        }
       },
       "limit": {
         "context": 1048576,
-        "output": 65536
+        "output": 131072
+      },
+      "knowledge": "2024-12"
+    }
+  },
+  {
+    "id": "xiaomi/mimo-v2.5",
+    "name": "Xiaomi: MiMo-V2.5",
+    "provider": "openrouter",
+    "family": "mimo",
+    "created_at": "2026-04-22 00:00:00 UTC",
+    "context_window": 1048576,
+    "max_output_tokens": 131072,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "audio",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "vision",
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.4,
+          "output_per_million": 2,
+          "cache_read_input_per_million": 0.08
+        }
       }
+    },
+    "metadata": {
+      "description": "MiMo-V2.5 is a native omnimodal model by Xiaomi. It delivers Pro-level agentic performance at roughly half the inference cost, while surpassing MiMo-V2-Omni in multimodal perception across image and video understanding...",
+      "architecture": {
+        "modality": "text+image+audio+video->text",
+        "input_modalities": [
+          "text",
+          "audio",
+          "image",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 131072,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
+      "source": "models.dev",
+      "provider_id": "openrouter",
+      "open_weights": true,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2026-04-22",
+      "interleaved": {
+        "field": "reasoning_details"
+      },
+      "cost": {
+        "input": 0.4,
+        "output": 2,
+        "cache_read": 0.08,
+        "context_over_200k": {
+          "input": 0.8,
+          "output": 4,
+          "cache_read": 0.16
+        }
+      },
+      "limit": {
+        "context": 1048576,
+        "output": 131072
+      },
+      "knowledge": "2024-12"
+    }
+  },
+  {
+    "id": "xiaomi/mimo-v2.5-pro",
+    "name": "Xiaomi: MiMo-V2.5-Pro",
+    "provider": "openrouter",
+    "family": "mimo",
+    "created_at": "2026-04-22 00:00:00 UTC",
+    "context_window": 1048576,
+    "max_output_tokens": 131072,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "streaming",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1,
+          "output_per_million": 3,
+          "cache_read_input_per_million": 0.2
+        }
+      }
+    },
+    "metadata": {
+      "description": "MiMo-V2.5-Pro is Xiaomi’s flagship model, delivering strong performance in general agentic capabilities, complex software engineering, and long-horizon tasks, with top rankings on benchmarks such as ClawEval, GDPVal, and SWE-bench Pro....",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 131072,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "response_format",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
+      "source": "models.dev",
+      "provider_id": "openrouter",
+      "open_weights": true,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2026-04-22",
+      "interleaved": {
+        "field": "reasoning_content"
+      },
+      "cost": {
+        "input": 1,
+        "output": 3,
+        "cache_read": 0.2,
+        "context_over_200k": {
+          "input": 2,
+          "output": 6,
+          "cache_read": 0.4
+        }
+      },
+      "limit": {
+        "context": 1048576,
+        "output": 131072
+      },
+      "knowledge": "2024-12"
+    }
+  },
+  {
+    "id": "z-ai/glm-4-32b",
+    "name": "Z.ai: GLM 4 32B ",
+    "provider": "openrouter",
+    "family": "z-ai",
+    "created_at": "2025-07-24 17:03:37 UTC",
+    "context_window": 128000,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.09999999999999999,
+          "output_per_million": 0.09999999999999999
+        }
+      }
+    },
+    "metadata": {
+      "description": "GLM 4 32B is a cost-effective foundation language model. It can efficiently perform complex tasks and has significantly enhanced capabilities in tool use, online search, and code-related intelligent tasks. It...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 128000,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "max_tokens",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
     }
   },
   {
@@ -22005,7 +56613,7 @@
     "name": "GLM 4.5",
     "provider": "openrouter",
     "family": "glm",
-    "created_at": "2025-07-28 00:00:00 +0530",
+    "created_at": "2025-07-28 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 96000,
     "knowledge_cutoff": null,
@@ -22020,7 +56628,8 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -22031,6 +56640,40 @@
       }
     },
     "metadata": {
+      "description": "GLM-4.5 is our latest flagship foundation model, purpose-built for agent-based applications. It leverages a Mixture-of-Experts (MoE) architecture and supports a context length of up to 128k tokens. GLM-4.5 delivers significantly...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 98304,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -22053,7 +56696,7 @@
     "name": "GLM 4.5 Air",
     "provider": "openrouter",
     "family": "glm-air",
-    "created_at": "2025-07-28 00:00:00 +0530",
+    "created_at": "2025-07-28 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 96000,
     "knowledge_cutoff": null,
@@ -22068,7 +56711,8 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -22079,6 +56723,39 @@
       }
     },
     "metadata": {
+      "description": "GLM-4.5-Air is the lightweight variant of our latest flagship model family, also purpose-built for agent-centric applications. Like GLM-4.5, it adopts the Mixture-of-Experts (MoE) architecture but with a more compact parameter...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 98304,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -22101,7 +56778,7 @@
     "name": "GLM 4.5 Air (free)",
     "provider": "openrouter",
     "family": "glm-air",
-    "created_at": "2025-07-28 00:00:00 +0530",
+    "created_at": "2025-07-28 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 96000,
     "knowledge_cutoff": null,
@@ -22114,10 +56791,39 @@
       ]
     },
     "capabilities": [
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "function_calling"
     ],
     "pricing": {},
     "metadata": {
+      "description": "GLM-4.5-Air is the lightweight variant of our latest flagship model family, also purpose-built for agent-centric applications. Like GLM-4.5, it adopts the Mixture-of-Experts (MoE) architecture but with a more compact parameter...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 96000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -22140,7 +56846,7 @@
     "name": "GLM 4.5V",
     "provider": "openrouter",
     "family": "glm",
-    "created_at": "2025-08-11 00:00:00 +0530",
+    "created_at": "2025-08-11 00:00:00 UTC",
     "context_window": 64000,
     "max_output_tokens": 16384,
     "knowledge_cutoff": null,
@@ -22158,7 +56864,8 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
@@ -22169,6 +56876,40 @@
       }
     },
     "metadata": {
+      "description": "GLM-4.5V is a vision-language foundation model for multimodal agent applications. Built on a Mixture-of-Experts (MoE) architecture with 106B parameters and 12B activated parameters, it achieves state-of-the-art results in video understanding,...",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 65536,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -22191,7 +56932,7 @@
     "name": "GLM 4.6",
     "provider": "openrouter",
     "family": "glm",
-    "created_at": "2025-09-30 00:00:00 +0530",
+    "created_at": "2025-09-30 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": null,
@@ -22206,18 +56947,57 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.6,
           "output_per_million": 2.2,
-          "cached_input_per_million": 0.11
+          "cache_read_input_per_million": 0.11
         }
       }
     },
     "metadata": {
+      "description": "Compared with GLM-4.5, this generation brings several key improvements: Longer context window: The context window has been expanded from 128K to 200K tokens, enabling the model to handle more complex...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 204800,
+        "max_completion_tokens": 204800,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -22241,7 +57021,7 @@
     "name": "GLM 4.6 (exacto)",
     "provider": "openrouter",
     "family": "glm",
-    "created_at": "2025-09-30 00:00:00 +0530",
+    "created_at": "2025-09-30 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 128000,
     "knowledge_cutoff": null,
@@ -22263,7 +57043,7 @@
         "standard": {
           "input_per_million": 0.6,
           "output_per_million": 1.9,
-          "cached_input_per_million": 0.11
+          "cache_read_input_per_million": 0.11
         }
       }
     },
@@ -22286,12 +57066,82 @@
       "knowledge": "2025-09"
     }
   },
+  {
+    "id": "z-ai/glm-4.6v",
+    "name": "Z.ai: GLM 4.6V",
+    "provider": "openrouter",
+    "family": "z-ai",
+    "created_at": "2025-12-08 15:24:22 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 24000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.3,
+          "output_per_million": 0.8999999999999999,
+          "cache_read_input_per_million": 0.049999999999999996
+        }
+      }
+    },
+    "metadata": {
+      "description": "GLM-4.6V is a large multimodal model designed for high-fidelity visual understanding and long-context reasoning across images, documents, and mixed media. It supports up to 128K tokens, processes complex page layouts...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 131072,
+        "max_completion_tokens": 24000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "max_tokens",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
   {
     "id": "z-ai/glm-4.7",
     "name": "GLM-4.7",
     "provider": "openrouter",
     "family": "glm",
-    "created_at": "2025-12-22 00:00:00 +0530",
+    "created_at": "2025-12-22 00:00:00 UTC",
     "context_window": 204800,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -22306,18 +57156,59 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.6,
           "output_per_million": 2.2,
-          "cached_input_per_million": 0.11
+          "cache_read_input_per_million": 0.11
         }
       }
     },
     "metadata": {
+      "description": "GLM-4.7 is Z.ai’s latest flagship model, featuring upgrades in two key areas: enhanced programming capabilities and more stable multi-step reasoning/execution. It demonstrates significant improvements in executing complex agent tasks while...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 202752,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -22344,7 +57235,7 @@
     "name": "GLM-4.7-Flash",
     "provider": "openrouter",
     "family": "glm",
-    "created_at": "2026-01-19 00:00:00 +0530",
+    "created_at": "2026-01-19 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 65535,
     "knowledge_cutoff": null,
@@ -22359,7 +57250,9 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
@@ -22370,6 +57263,43 @@
       }
     },
     "metadata": {
+      "description": "As a 30B-class SOTA model, GLM-4.7-Flash offers a new option that balances performance and efficiency. It is further optimized for agentic coding use cases, strengthening coding capabilities, long-horizon task planning,...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 202752,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -22394,7 +57324,7 @@
     "name": "GLM-5",
     "provider": "openrouter",
     "family": "glm",
-    "created_at": "2026-02-12 00:00:00 +0530",
+    "created_at": "2026-02-12 00:00:00 UTC",
     "context_window": 202752,
     "max_output_tokens": 131000,
     "knowledge_cutoff": null,
@@ -22409,18 +57339,59 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1,
           "output_per_million": 3.2,
-          "cached_input_per_million": 0.2
+          "cache_read_input_per_million": 0.2
         }
       }
     },
     "metadata": {
+      "description": "GLM-5 is Z.ai’s flagship open-source foundation model engineered for complex systems design and long-horizon agent workflows. Built for expert developers, it delivers production-grade performance on large-scale programming tasks, rivaling leading...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 202752,
+        "max_completion_tokens": null,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -22446,7 +57417,7 @@
     "name": "GLM-5-Turbo",
     "provider": "openrouter",
     "family": "glm",
-    "created_at": "2026-03-16 00:00:00 +0530",
+    "created_at": "2026-03-16 00:00:00 UTC",
     "context_window": 202752,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -22461,18 +57432,56 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.96,
           "output_per_million": 3.2,
-          "cached_input_per_million": 0.192
+          "cache_read_input_per_million": 0.192
         }
       }
     },
     "metadata": {
+      "description": "GLM-5 Turbo is a new model from Z.ai designed for fast inference and strong performance in agent-driven environments such as OpenClaw scenarios. It is deeply optimized for real-world agent workflows...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 202752,
+        "max_completion_tokens": 131072,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "max_tokens",
+        "min_p",
+        "presence_penalty",
+        "reasoning",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": false,
@@ -22499,7 +57508,7 @@
     "name": "GLM-5.1",
     "provider": "openrouter",
     "family": "glm",
-    "created_at": "2026-04-07 00:00:00 +0530",
+    "created_at": "2026-04-07 00:00:00 UTC",
     "context_window": 202752,
     "max_output_tokens": 131072,
     "knowledge_cutoff": null,
@@ -22514,18 +57523,61 @@
     "capabilities": [
       "function_calling",
       "structured_output",
-      "reasoning"
+      "reasoning",
+      "streaming",
+      "predicted_outputs"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.4,
           "output_per_million": 4.4,
-          "cached_input_per_million": 0.26
+          "cache_read_input_per_million": 0.26
         }
       }
     },
     "metadata": {
+      "description": "GLM-5.1 delivers a major leap in coding capability, with particularly significant gains in handling long-horizon tasks. Unlike previous models built around minute-level interactions, GLM-5.1 can work independently and continuously on...",
+      "architecture": {
+        "modality": "text->text",
+        "input_modalities": [
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 202752,
+        "max_completion_tokens": 65535,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "parallel_tool_calls",
+        "presence_penalty",
+        "reasoning",
+        "reasoning_effort",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ],
       "source": "models.dev",
       "provider_id": "openrouter",
       "open_weights": true,
@@ -22546,12 +57598,639 @@
       }
     }
   },
+  {
+    "id": "z-ai/glm-5v-turbo",
+    "name": "Z.ai: GLM 5V Turbo",
+    "provider": "openrouter",
+    "family": "z-ai",
+    "created_at": "2026-04-01 16:37:38 UTC",
+    "context_window": 202752,
+    "max_output_tokens": 131072,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.2,
+          "output_per_million": 4.0,
+          "cache_read_input_per_million": 0.24
+        }
+      }
+    },
+    "metadata": {
+      "description": "GLM-5V-Turbo is Z.ai’s first native multimodal agent foundation model, built for vision-based coding and agent-driven tasks. It natively handles image, video, and text inputs, excels at long-horizon planning, complex coding,...",
+      "architecture": {
+        "modality": "text+image+video->text",
+        "input_modalities": [
+          "image",
+          "text",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Other",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 202752,
+        "max_completion_tokens": 131072,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "~anthropic/claude-haiku-latest",
+    "name": "Anthropic Claude Haiku Latest",
+    "provider": "openrouter",
+    "family": "~anthropic",
+    "created_at": "2026-04-27 19:34:52 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 64000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "image",
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.0,
+          "output_per_million": 5.0,
+          "cache_read_input_per_million": 0.09999999999999999
+        }
+      }
+    },
+    "metadata": {
+      "description": "This model always redirects to the latest model in the Anthropic Claude Haiku family.",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Router",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 200000,
+        "max_completion_tokens": 64000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "~anthropic/claude-opus-latest",
+    "name": "Anthropic: Claude Opus Latest",
+    "provider": "openrouter",
+    "family": "~anthropic",
+    "created_at": "2026-04-21 18:16:01 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5.0,
+          "output_per_million": 25.0,
+          "cache_read_input_per_million": 0.5
+        }
+      }
+    },
+    "metadata": {
+      "description": "This model always redirects to the latest model in the Claude Opus family.",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Router",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 128000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "tool_choice",
+        "tools",
+        "verbosity"
+      ]
+    }
+  },
+  {
+    "id": "~anthropic/claude-sonnet-latest",
+    "name": "Anthropic Claude Sonnet Latest",
+    "provider": "openrouter",
+    "family": "~anthropic",
+    "created_at": "2026-04-27 19:32:48 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 3.0,
+          "output_per_million": 15.0,
+          "cache_read_input_per_million": 0.3
+        }
+      }
+    },
+    "metadata": {
+      "description": "This model always redirects to the latest model in the Anthropic Claude Sonnet family.",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Router",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1000000,
+        "max_completion_tokens": 128000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_p",
+        "verbosity"
+      ]
+    }
+  },
+  {
+    "id": "~google/gemini-flash-latest",
+    "name": "Google Gemini Flash Latest",
+    "provider": "openrouter",
+    "family": "~google",
+    "created_at": "2026-04-27 19:33:18 UTC",
+    "context_window": 1048576,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "file",
+        "audio",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.5,
+          "output_per_million": 3.0,
+          "cache_read_input_per_million": 0.049999999999999996,
+          "reasoning_output_per_million": 3.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "This model always redirects to the latest model in the Google Gemini Flash family.",
+      "architecture": {
+        "modality": "text+image+file+audio+video->text",
+        "input_modalities": [
+          "text",
+          "image",
+          "file",
+          "audio",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Router",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "~google/gemini-pro-latest",
+    "name": "Google Gemini Pro Latest",
+    "provider": "openrouter",
+    "family": "~google",
+    "created_at": "2026-04-27 19:34:11 UTC",
+    "context_window": 1048576,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "audio",
+        "file",
+        "image",
+        "text",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 2.0,
+          "output_per_million": 12.0,
+          "cache_read_input_per_million": 0.19999999999999998,
+          "reasoning_output_per_million": 12.0
+        }
+      }
+    },
+    "metadata": {
+      "description": "This model always redirects to the latest model in the Google Gemini Pro family.",
+      "architecture": {
+        "modality": "text+image+file+audio+video->text",
+        "input_modalities": [
+          "audio",
+          "file",
+          "image",
+          "text",
+          "video"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Router",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1048576,
+        "max_completion_tokens": 65536,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "~moonshotai/kimi-latest",
+    "name": "MoonshotAI Kimi Latest",
+    "provider": "openrouter",
+    "family": "~moonshotai",
+    "created_at": "2026-04-27 19:33:48 UTC",
+    "context_window": 262144,
+    "max_output_tokens": 16384,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "predicted_outputs"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.75,
+          "output_per_million": 3.5,
+          "cache_read_input_per_million": 0.15
+        }
+      }
+    },
+    "metadata": {
+      "description": "This model always redirects to the latest model in the MoonshotAI Kimi family.",
+      "architecture": {
+        "modality": "text+image->text",
+        "input_modalities": [
+          "text",
+          "image"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Router",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 262144,
+        "max_completion_tokens": 16384,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "frequency_penalty",
+        "include_reasoning",
+        "logit_bias",
+        "logprobs",
+        "max_tokens",
+        "min_p",
+        "parallel_tool_calls",
+        "presence_penalty",
+        "reasoning",
+        "reasoning_effort",
+        "repetition_penalty",
+        "response_format",
+        "seed",
+        "stop",
+        "structured_outputs",
+        "temperature",
+        "tool_choice",
+        "tools",
+        "top_k",
+        "top_logprobs",
+        "top_p"
+      ]
+    }
+  },
+  {
+    "id": "~openai/gpt-latest",
+    "name": "OpenAI GPT Latest",
+    "provider": "openrouter",
+    "family": "~openai",
+    "created_at": "2026-04-27 19:32:14 UTC",
+    "context_window": 1050000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "file",
+        "image",
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5.0,
+          "output_per_million": 30.0,
+          "cache_read_input_per_million": 0.5
+        }
+      }
+    },
+    "metadata": {
+      "description": "This model always redirects to the latest model in the OpenAI GPT family.",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "file",
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Router",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 1050000,
+        "max_completion_tokens": 128000,
+        "is_moderated": true
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ]
+    }
+  },
+  {
+    "id": "~openai/gpt-mini-latest",
+    "name": "OpenAI GPT Mini Latest",
+    "provider": "openrouter",
+    "family": "~openai",
+    "created_at": "2026-04-27 19:34:31 UTC",
+    "context_window": 400000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "file",
+        "image",
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.75,
+          "output_per_million": 4.5,
+          "cache_read_input_per_million": 0.075
+        }
+      }
+    },
+    "metadata": {
+      "description": "This model always redirects to the latest model in the OpenAI GPT Mini family.",
+      "architecture": {
+        "modality": "text+image+file->text",
+        "input_modalities": [
+          "file",
+          "image",
+          "text"
+        ],
+        "output_modalities": [
+          "text"
+        ],
+        "tokenizer": "Router",
+        "instruct_type": null
+      },
+      "top_provider": {
+        "context_length": 400000,
+        "max_completion_tokens": 128000,
+        "is_moderated": false
+      },
+      "per_request_limits": null,
+      "supported_parameters": [
+        "include_reasoning",
+        "max_completion_tokens",
+        "max_tokens",
+        "reasoning",
+        "response_format",
+        "seed",
+        "structured_outputs",
+        "tool_choice",
+        "tools"
+      ]
+    }
+  },
   {
     "id": "sonar",
     "name": "Sonar",
     "provider": "perplexity",
     "family": "sonar",
-    "created_at": "2024-01-01 00:00:00 +0530",
+    "created_at": "2024-01-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": "2025-09-01",
@@ -22563,7 +58242,9 @@
         "text"
       ]
     },
-    "capabilities": [],
+    "capabilities": [
+      "vision"
+    ],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -22595,7 +58276,7 @@
     "name": "Perplexity Sonar Deep Research",
     "provider": "perplexity",
     "family": null,
-    "created_at": "2025-02-01 00:00:00 +0530",
+    "created_at": "2025-02-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 32768,
     "knowledge_cutoff": null,
@@ -22643,7 +58324,7 @@
     "name": "Sonar Pro",
     "provider": "perplexity",
     "family": "sonar-pro",
-    "created_at": "2024-01-01 00:00:00 +0530",
+    "created_at": "2024-01-01 00:00:00 UTC",
     "context_window": 200000,
     "max_output_tokens": 8192,
     "knowledge_cutoff": "2025-09-01",
@@ -22685,12 +58366,39 @@
       "knowledge": "2025-09-01"
     }
   },
+  {
+    "id": "sonar-reasoning",
+    "name": "sonar-reasoning",
+    "provider": "perplexity",
+    "family": null,
+    "created_at": null,
+    "context_window": 128000,
+    "max_output_tokens": 4096,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "vision",
+      "reasoning"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.0,
+          "output_per_million": 5.0
+        }
+      }
+    },
+    "metadata": {}
+  },
   {
     "id": "sonar-reasoning-pro",
     "name": "Sonar Reasoning Pro",
     "provider": "perplexity",
     "family": "sonar-reasoning",
-    "created_at": "2024-01-01 00:00:00 +0530",
+    "created_at": "2024-01-01 00:00:00 UTC",
     "context_window": 128000,
     "max_output_tokens": 4096,
     "knowledge_cutoff": "2025-09-01",
@@ -22734,17 +58442,18 @@
     }
   },
   {
-    "id": "deepseek-ai/deepseek-v3.1-maas",
-    "name": "DeepSeek V3.1",
+    "id": "claude-3-5-haiku@20241022",
+    "name": "Claude Haiku 3.5",
     "provider": "vertexai",
-    "family": "deepseek",
-    "created_at": "2025-08-28 00:00:00 +0530",
-    "context_window": 163840,
-    "max_output_tokens": 32768,
-    "knowledge_cutoff": null,
+    "family": "claude-haiku",
+    "created_at": "2024-10-22 00:00:00 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": "2024-07-31",
     "modalities": {
       "input": [
         "text",
+        "image",
         "pdf"
       ],
       "output": [
@@ -22753,47 +58462,51 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output",
-      "reasoning",
       "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
-          "input_per_million": 0.6,
-          "output_per_million": 1.7
+          "input_per_million": 0.8,
+          "output_per_million": 4,
+          "cache_read_input_per_million": 0.08,
+          "cache_write_input_per_million": 1
         }
       }
     },
     "metadata": {
       "source": "models.dev",
       "provider_id": "google-vertex",
-      "open_weights": true,
-      "attachment": false,
+      "open_weights": false,
+      "attachment": true,
       "temperature": true,
-      "last_updated": "2025-08-28",
+      "last_updated": "2024-10-22",
       "cost": {
-        "input": 0.6,
-        "output": 1.7
+        "input": 0.8,
+        "output": 4,
+        "cache_read": 0.08,
+        "cache_write": 1
       },
       "limit": {
-        "context": 163840,
-        "output": 32768
-      }
+        "context": 200000,
+        "output": 8192
+      },
+      "knowledge": "2024-07-31"
     }
   },
   {
-    "id": "deepseek-ai/deepseek-v3.2-maas",
-    "name": "DeepSeek V3.2",
+    "id": "claude-3-5-sonnet@20241022",
+    "name": "Claude Sonnet 3.5 v2",
     "provider": "vertexai",
-    "family": "deepseek",
-    "created_at": "2025-12-17 00:00:00 +0530",
-    "context_window": 163840,
-    "max_output_tokens": 65536,
-    "knowledge_cutoff": null,
+    "family": "claude-sonnet",
+    "created_at": "2024-10-22 00:00:00 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": "2024-04-30",
     "modalities": {
       "input": [
         "text",
+        "image",
         "pdf"
       ],
       "output": [
@@ -22802,35 +58515,862 @@
     },
     "capabilities": [
       "function_calling",
-      "structured_output",
-      "reasoning",
       "vision"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
-          "input_per_million": 0.56,
-          "output_per_million": 1.68,
-          "cached_input_per_million": 0.056
+          "input_per_million": 3,
+          "output_per_million": 15,
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
         }
       }
     },
     "metadata": {
       "source": "models.dev",
       "provider_id": "google-vertex",
-      "open_weights": true,
-      "attachment": false,
+      "open_weights": false,
+      "attachment": true,
       "temperature": true,
-      "last_updated": "2026-04-04",
+      "last_updated": "2024-10-22",
       "cost": {
-        "input": 0.56,
-        "output": 1.68,
-        "cache_read": 0.056
+        "input": 3,
+        "output": 15,
+        "cache_read": 0.3,
+        "cache_write": 3.75
       },
       "limit": {
-        "context": 163840,
-        "output": 65536
+        "context": 200000,
+        "output": 8192
+      },
+      "knowledge": "2024-04-30"
+    }
+  },
+  {
+    "id": "claude-3-7-sonnet@20250219",
+    "name": "Claude Sonnet 3.7",
+    "provider": "vertexai",
+    "family": "claude-sonnet",
+    "created_at": "2025-02-19 00:00:00 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 64000,
+    "knowledge_cutoff": "2024-10-31",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 3,
+          "output_per_million": 15,
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
+        }
       }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "google-vertex",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2025-02-19",
+      "cost": {
+        "input": 3,
+        "output": 15,
+        "cache_read": 0.3,
+        "cache_write": 3.75
+      },
+      "limit": {
+        "context": 200000,
+        "output": 64000
+      },
+      "knowledge": "2024-10-31"
+    }
+  },
+  {
+    "id": "claude-haiku-4-5@20251001",
+    "name": "Claude Haiku 4.5",
+    "provider": "vertexai",
+    "family": "claude-haiku",
+    "created_at": "2025-10-15 00:00:00 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 64000,
+    "knowledge_cutoff": "2025-02-28",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1,
+          "output_per_million": 5,
+          "cache_read_input_per_million": 0.1,
+          "cache_write_input_per_million": 1.25
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "google-vertex",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2025-10-15",
+      "cost": {
+        "input": 1,
+        "output": 5,
+        "cache_read": 0.1,
+        "cache_write": 1.25
+      },
+      "limit": {
+        "context": 200000,
+        "output": 64000
+      },
+      "knowledge": "2025-02-28"
+    }
+  },
+  {
+    "id": "claude-opus-4-1@20250805",
+    "name": "Claude Opus 4.1",
+    "provider": "vertexai",
+    "family": "claude-opus",
+    "created_at": "2025-08-05 00:00:00 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 32000,
+    "knowledge_cutoff": "2025-03-31",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 15,
+          "output_per_million": 75,
+          "cache_read_input_per_million": 1.5,
+          "cache_write_input_per_million": 18.75
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "google-vertex",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2025-08-05",
+      "cost": {
+        "input": 15,
+        "output": 75,
+        "cache_read": 1.5,
+        "cache_write": 18.75
+      },
+      "limit": {
+        "context": 200000,
+        "output": 32000
+      },
+      "knowledge": "2025-03-31"
+    }
+  },
+  {
+    "id": "claude-opus-4-5@20251101",
+    "name": "Claude Opus 4.5",
+    "provider": "vertexai",
+    "family": "claude-opus",
+    "created_at": "2025-11-01 00:00:00 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 64000,
+    "knowledge_cutoff": "2025-03-31",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5,
+          "output_per_million": 25,
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "google-vertex",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2025-11-01",
+      "cost": {
+        "input": 5,
+        "output": 25,
+        "cache_read": 0.5,
+        "cache_write": 6.25
+      },
+      "limit": {
+        "context": 200000,
+        "output": 64000
+      },
+      "knowledge": "2025-03-31"
+    }
+  },
+  {
+    "id": "claude-opus-4-6@default",
+    "name": "Claude Opus 4.6",
+    "provider": "vertexai",
+    "family": "claude-opus",
+    "created_at": "2026-02-05 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": "2025-05-31",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5,
+          "output_per_million": 25,
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "google-vertex",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2026-03-13",
+      "cost": {
+        "input": 5,
+        "output": 25,
+        "cache_read": 0.5,
+        "cache_write": 6.25,
+        "context_over_200k": {
+          "input": 10,
+          "output": 37.5,
+          "cache_read": 1,
+          "cache_write": 12.5
+        }
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 128000
+      },
+      "knowledge": "2025-05-31"
+    }
+  },
+  {
+    "id": "claude-opus-4-7@default",
+    "name": "Claude Opus 4.7",
+    "provider": "vertexai",
+    "family": "claude-opus",
+    "created_at": "2026-04-16 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 128000,
+    "knowledge_cutoff": "2026-01-31",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 5,
+          "output_per_million": 25,
+          "cache_read_input_per_million": 0.5,
+          "cache_write_input_per_million": 6.25
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "google-vertex",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": false,
+      "last_updated": "2026-04-16",
+      "cost": {
+        "input": 5,
+        "output": 25,
+        "cache_read": 0.5,
+        "cache_write": 6.25,
+        "context_over_200k": {
+          "input": 10,
+          "output": 37.5,
+          "cache_read": 1,
+          "cache_write": 12.5
+        }
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 128000
+      },
+      "knowledge": "2026-01-31"
+    }
+  },
+  {
+    "id": "claude-opus-4@20250514",
+    "name": "Claude Opus 4",
+    "provider": "vertexai",
+    "family": "claude-opus",
+    "created_at": "2025-05-22 00:00:00 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 32000,
+    "knowledge_cutoff": "2025-03-31",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 15,
+          "output_per_million": 75,
+          "cache_read_input_per_million": 1.5,
+          "cache_write_input_per_million": 18.75
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "google-vertex",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2025-05-22",
+      "cost": {
+        "input": 15,
+        "output": 75,
+        "cache_read": 1.5,
+        "cache_write": 18.75
+      },
+      "limit": {
+        "context": 200000,
+        "output": 32000
+      },
+      "knowledge": "2025-03-31"
+    }
+  },
+  {
+    "id": "claude-sonnet-4-5@20250929",
+    "name": "Claude Sonnet 4.5",
+    "provider": "vertexai",
+    "family": "claude-sonnet",
+    "created_at": "2025-09-29 00:00:00 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 64000,
+    "knowledge_cutoff": "2025-07-31",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 3,
+          "output_per_million": 15,
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "google-vertex",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2025-09-29",
+      "cost": {
+        "input": 3,
+        "output": 15,
+        "cache_read": 0.3,
+        "cache_write": 3.75
+      },
+      "limit": {
+        "context": 200000,
+        "output": 64000
+      },
+      "knowledge": "2025-07-31"
+    }
+  },
+  {
+    "id": "claude-sonnet-4-6@default",
+    "name": "Claude Sonnet 4.6",
+    "provider": "vertexai",
+    "family": "claude-sonnet",
+    "created_at": "2026-02-17 00:00:00 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 64000,
+    "knowledge_cutoff": "2025-08-31",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 3,
+          "output_per_million": 15,
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "google-vertex",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2026-03-13",
+      "cost": {
+        "input": 3,
+        "output": 15,
+        "cache_read": 0.3,
+        "cache_write": 3.75,
+        "context_over_200k": {
+          "input": 6,
+          "output": 22.5,
+          "cache_read": 0.6,
+          "cache_write": 7.5
+        }
+      },
+      "limit": {
+        "context": 200000,
+        "output": 64000
+      },
+      "knowledge": "2025-08-31"
+    }
+  },
+  {
+    "id": "claude-sonnet-4@20250514",
+    "name": "Claude Sonnet 4",
+    "provider": "vertexai",
+    "family": "claude-sonnet",
+    "created_at": "2025-05-22 00:00:00 UTC",
+    "context_window": 200000,
+    "max_output_tokens": 64000,
+    "knowledge_cutoff": "2025-03-31",
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 3,
+          "output_per_million": 15,
+          "cache_read_input_per_million": 0.3,
+          "cache_write_input_per_million": 3.75
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "google-vertex",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2025-05-22",
+      "cost": {
+        "input": 3,
+        "output": 15,
+        "cache_read": 0.3,
+        "cache_write": 3.75
+      },
+      "limit": {
+        "context": 200000,
+        "output": 64000
+      },
+      "knowledge": "2025-03-31"
+    }
+  },
+  {
+    "id": "gemini-1.5-flash",
+    "name": "Gemini 1.5 Flash",
+    "provider": "vertexai",
+    "family": "gemini-flash",
+    "created_at": "2024-05-14 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "audio",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.075,
+          "output_per_million": 0.3,
+          "cache_read_input_per_million": 0.01875
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "google",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2024-05-14",
+      "cost": {
+        "input": 0.075,
+        "output": 0.3,
+        "cache_read": 0.01875
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 8192
+      },
+      "knowledge": "2024-04"
+    }
+  },
+  {
+    "id": "gemini-1.5-flash-002",
+    "name": "gemini-1.5-flash-002",
+    "provider": "vertexai",
+    "family": "gemini-1.5",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "source": "known_models"
+    }
+  },
+  {
+    "id": "gemini-1.5-flash-8b",
+    "name": "Gemini 1.5 Flash-8B",
+    "provider": "vertexai",
+    "family": "gemini-flash",
+    "created_at": "2024-10-03 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "audio",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.0375,
+          "output_per_million": 0.15,
+          "cache_read_input_per_million": 0.01
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "google",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2024-10-03",
+      "cost": {
+        "input": 0.0375,
+        "output": 0.15,
+        "cache_read": 0.01
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 8192
+      },
+      "knowledge": "2024-04"
+    }
+  },
+  {
+    "id": "gemini-1.5-pro",
+    "name": "Gemini 1.5 Pro",
+    "provider": "vertexai",
+    "family": "gemini-pro",
+    "created_at": "2024-02-15 00:00:00 UTC",
+    "context_window": 1000000,
+    "max_output_tokens": 8192,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "audio",
+        "video"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 1.25,
+          "output_per_million": 5,
+          "cache_read_input_per_million": 0.3125
+        }
+      }
+    },
+    "metadata": {
+      "source": "models.dev",
+      "provider_id": "google",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2024-02-15",
+      "cost": {
+        "input": 1.25,
+        "output": 5,
+        "cache_read": 0.3125
+      },
+      "limit": {
+        "context": 1000000,
+        "output": 8192
+      },
+      "knowledge": "2024-04"
+    }
+  },
+  {
+    "id": "gemini-1.5-pro-002",
+    "name": "gemini-1.5-pro-002",
+    "provider": "vertexai",
+    "family": "gemini-1.5",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "version_id": "default",
+      "open_source_category": "PROPRIETARY",
+      "launch_stage": "GA",
+      "supported_actions": {
+        "openNotebook": {
+          "references": {
+            "us-central1": {
+              "uri": "https://colab.research.google.com/github/GoogleCloudPlatform/generative-ai/blob/main/gemini/use-cases/retail/product_attributes_extraction.ipynb"
+            }
+          },
+          "title": "Open Notebook",
+          "resourceTitle": "Notebook",
+          "resourceUseCase": "Product Attributes Extraction",
+          "resourceDescription": "Extract product descriptions and attribute json from images using Gemini 1.5 Pro. This notebook also shows the use of self-correcting prompt to improve the quality of the output."
+        },
+        "openGenerationAiStudio": {
+          "references": {
+            "us-central1": {
+              "uri": "https://console.cloud.google.com/vertex-ai/studio/freeform?model=gemini-1.5-pro-002"
+            }
+          }
+        },
+        "openEvaluationPipeline": {
+          "references": {
+            "us-central1": {
+              "uri": "https://console.cloud.google.com/vertex-ai/pipelines/vertex-ai-templates/autosxs-template"
+            }
+          },
+          "title": "Evaluate"
+        },
+        "openNotebooks": {
+          "notebooks": [
+            {
+              "references": {
+                "us-central1": {
+                  "uri": "https://colab.research.google.com/github/GoogleCloudPlatform/generative-ai/blob/main/gemini/getting-started/intro_gemini_1_5_pro.ipynb"
+                }
+              },
+              "title": "Open Notebook"
+            },
+            {
+              "references": {
+                "us-central1": {
+                  "uri": "https://colab.research.google.com/github/GoogleCloudPlatform/generative-ai/blob/main/gemini/getting-started/intro_gemini_1_5_pro.ipynb"
+                }
+              },
+              "title": "Open Notebook",
+              "resourceTitle": "Notebook",
+              "resourceUseCase": "Vertex AI Gemini API 1.5 Pro",
+              "resourceDescription": "Use the Vertex AI Gemini API 1.5 Pro model to process images, video, audio, and text simultaneously."
+            },
+            {
+              "references": {
+                "us-central1": {
+                  "uri": "https://colab.research.google.com/github/GoogleCloudPlatform/generative-ai/blob/main/gemini/use-cases/retail/product_attributes_extraction.ipynb"
+                }
+              },
+              "title": "Open Notebook",
+              "resourceTitle": "Notebook",
+              "resourceUseCase": "Product Attributes Extraction",
+              "resourceDescription": "Extract product descriptions and attribute json from images using Gemini 1.5 Pro. This notebook also shows the use of self-correcting prompt to improve the quality of the output."
+            }
+          ]
+        }
+      },
+      "publisher_model_template": "projects/{project}/locations/{location}/publishers/google/models/gemini-1.5-pro-002@default"
     }
   },
   {
@@ -22838,7 +59378,7 @@
     "name": "Gemini 2.0 Flash",
     "provider": "vertexai",
     "family": "gemini-flash",
-    "created_at": "2024-12-11 00:00:00 +0530",
+    "created_at": "2024-12-11 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -22856,18 +59396,56 @@
     },
     "capabilities": [
       "function_calling",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.15,
           "output_per_million": 0.6,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       }
     },
     "metadata": {
+      "version_id": "default",
+      "open_source_category": null,
+      "launch_stage": "GA",
+      "supported_actions": {
+        "openNotebook": {
+          "references": {
+            "us-central1": {
+              "uri": "https://colab.research.google.com/github/GoogleCloudPlatform/generative-ai/blob/main/gemini/getting-started/intro_gemini_2_0_flash.ipynb"
+            }
+          },
+          "resourceTitle": "Notebook",
+          "resourceUseCase": "Vertex Serving",
+          "resourceDescription": "Intro to Gemini 2.0 Flash."
+        },
+        "openGenerationAiStudio": {
+          "references": {
+            "us-central1": {
+              "uri": "https://console.cloud.google.com/vertex-ai/generative/multimodal/create/text?model=gemini-2.0-flash-001"
+            }
+          }
+        },
+        "openNotebooks": {
+          "notebooks": [
+            {
+              "references": {
+                "us-central1": {
+                  "uri": "https://colab.research.google.com/github/GoogleCloudPlatform/generative-ai/blob/main/gemini/getting-started/intro_gemini_2_0_flash.ipynb"
+                }
+              },
+              "resourceTitle": "Notebook",
+              "resourceUseCase": "Vertex Serving",
+              "resourceDescription": "Intro to Gemini 2.0 Flash."
+            }
+          ]
+        }
+      },
+      "publisher_model_template": "projects/{project}/locations/{location}/publishers/google/models/gemini-2.0-flash@default",
       "source": "models.dev",
       "provider_id": "google-vertex",
       "open_weights": false,
@@ -22886,12 +59464,92 @@
       "knowledge": "2024-06"
     }
   },
+  {
+    "id": "gemini-2.0-flash-001",
+    "name": "gemini-2.0-flash-001",
+    "provider": "vertexai",
+    "family": "gemini-2",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "version_id": "default",
+      "open_source_category": null,
+      "launch_stage": "GA",
+      "supported_actions": {
+        "openNotebook": {
+          "references": {
+            "us-central1": {
+              "uri": "https://colab.research.google.com/github/GoogleCloudPlatform/generative-ai/blob/main/gemini/getting-started/intro_gemini_2_0_flash.ipynb"
+            }
+          },
+          "resourceTitle": "Notebook",
+          "resourceUseCase": "Vertex Serving",
+          "resourceDescription": "Intro to Gemini 2.0 Flash."
+        },
+        "openGenerationAiStudio": {
+          "references": {
+            "us-central1": {
+              "uri": "https://console.cloud.google.com/vertex-ai/generative/multimodal/create/text?model=gemini-2.0-flash-001"
+            }
+          }
+        },
+        "openNotebooks": {
+          "notebooks": [
+            {
+              "references": {
+                "us-central1": {
+                  "uri": "https://colab.research.google.com/github/GoogleCloudPlatform/generative-ai/blob/main/gemini/getting-started/intro_gemini_2_0_flash.ipynb"
+                }
+              },
+              "resourceTitle": "Notebook",
+              "resourceUseCase": "Vertex Serving",
+              "resourceDescription": "Intro to Gemini 2.0 Flash."
+            }
+          ]
+        }
+      },
+      "publisher_model_template": "projects/{project}/locations/{location}/publishers/google/models/gemini-2.0-flash-001@default"
+    }
+  },
+  {
+    "id": "gemini-2.0-flash-exp",
+    "name": "gemini-2.0-flash-exp",
+    "provider": "vertexai",
+    "family": "gemini-2",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "source": "known_models"
+    }
+  },
   {
     "id": "gemini-2.0-flash-lite",
     "name": "Gemini 2.0 Flash Lite",
     "provider": "vertexai",
     "family": "gemini-flash-lite",
-    "created_at": "2024-12-11 00:00:00 +0530",
+    "created_at": "2024-12-11 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 8192,
     "knowledge_cutoff": null,
@@ -22937,12 +59595,38 @@
       "knowledge": "2024-06"
     }
   },
+  {
+    "id": "gemini-2.0-flash-lite-001",
+    "name": "gemini-2.0-flash-lite-001",
+    "provider": "vertexai",
+    "family": "gemini-2",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "version_id": "default",
+      "open_source_category": null,
+      "launch_stage": "GA",
+      "supported_actions": null,
+      "publisher_model_template": "projects/{project}/locations/{location}/publishers/google/models/gemini-2.0-flash-lite-001@default"
+    }
+  },
   {
     "id": "gemini-2.5-flash",
     "name": "Gemini 2.5 Flash",
     "provider": "vertexai",
     "family": "gemini-flash",
-    "created_at": "2025-06-17 00:00:00 +0530",
+    "created_at": "2025-06-17 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -22961,18 +59645,25 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.3,
           "output_per_million": 2.5,
-          "cached_input_per_million": 0.075
+          "cache_read_input_per_million": 0.075,
+          "cache_write_input_per_million": 0.383
         }
       }
     },
     "metadata": {
+      "version_id": "default",
+      "open_source_category": null,
+      "launch_stage": "GA",
+      "supported_actions": null,
+      "publisher_model_template": "projects/{project}/locations/{location}/publishers/google/models/gemini-2.5-flash@default",
       "source": "models.dev",
       "provider_id": "google-vertex",
       "open_weights": false,
@@ -22997,7 +59688,7 @@
     "name": "Gemini 2.5 Flash Lite",
     "provider": "vertexai",
     "family": "gemini-flash-lite",
-    "created_at": "2025-06-17 00:00:00 +0530",
+    "created_at": "2025-06-17 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -23016,18 +59707,24 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.1,
           "output_per_million": 0.4,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       }
     },
     "metadata": {
+      "version_id": "default",
+      "open_source_category": null,
+      "launch_stage": "GA",
+      "supported_actions": null,
+      "publisher_model_template": "projects/{project}/locations/{location}/publishers/google/models/gemini-2.5-flash-lite@default",
       "source": "models.dev",
       "provider_id": "google-vertex",
       "open_weights": false,
@@ -23051,7 +59748,7 @@
     "name": "Gemini 2.5 Flash Lite Preview 06-17",
     "provider": "vertexai",
     "family": "gemini-flash-lite",
-    "created_at": "2025-06-17 00:00:00 +0530",
+    "created_at": "2025-06-17 00:00:00 UTC",
     "context_window": 65536,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -23077,7 +59774,7 @@
         "standard": {
           "input_per_million": 0.1,
           "output_per_million": 0.4,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       }
     },
@@ -23105,7 +59802,7 @@
     "name": "Gemini 2.5 Flash Lite Preview 09-25",
     "provider": "vertexai",
     "family": "gemini-flash-lite",
-    "created_at": "2025-09-25 00:00:00 +0530",
+    "created_at": "2025-09-25 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -23131,7 +59828,7 @@
         "standard": {
           "input_per_million": 0.1,
           "output_per_million": 0.4,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       }
     },
@@ -23159,7 +59856,7 @@
     "name": "Gemini 2.5 Flash Preview 04-17",
     "provider": "vertexai",
     "family": "gemini-flash",
-    "created_at": "2025-04-17 00:00:00 +0530",
+    "created_at": "2025-04-17 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -23178,18 +59875,32 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.15,
           "output_per_million": 0.6,
-          "cached_input_per_million": 0.0375
+          "cache_read_input_per_million": 0.0375
         }
       }
     },
     "metadata": {
+      "version_id": "default",
+      "open_source_category": null,
+      "launch_stage": "PUBLIC_PREVIEW",
+      "supported_actions": {
+        "openGenerationAiStudio": {
+          "references": {
+            "us-central1": {
+              "uri": "https://console.cloud.google.com/vertex-ai/generative/multimodal/create/text?model=gemini-2.5-flash-preview-04-17"
+            }
+          }
+        }
+      },
+      "publisher_model_template": "projects/{project}/locations/{location}/publishers/google/models/gemini-2.5-flash-preview-04-17@default",
       "source": "models.dev",
       "provider_id": "google-vertex",
       "open_weights": false,
@@ -23213,7 +59924,7 @@
     "name": "Gemini 2.5 Flash Preview 05-20",
     "provider": "vertexai",
     "family": "gemini-flash",
-    "created_at": "2025-05-20 00:00:00 +0530",
+    "created_at": "2025-05-20 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -23239,7 +59950,7 @@
         "standard": {
           "input_per_million": 0.15,
           "output_per_million": 0.6,
-          "cached_input_per_million": 0.0375
+          "cache_read_input_per_million": 0.0375
         }
       }
     },
@@ -23267,7 +59978,7 @@
     "name": "Gemini 2.5 Flash Preview 09-25",
     "provider": "vertexai",
     "family": "gemini-flash",
-    "created_at": "2025-09-25 00:00:00 +0530",
+    "created_at": "2025-09-25 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -23293,7 +60004,8 @@
         "standard": {
           "input_per_million": 0.3,
           "output_per_million": 2.5,
-          "cached_input_per_million": 0.075
+          "cache_read_input_per_million": 0.075,
+          "cache_write_input_per_million": 0.383
         }
       }
     },
@@ -23317,12 +60029,38 @@
       "knowledge": "2025-01"
     }
   },
+  {
+    "id": "gemini-2.5-flash-tts",
+    "name": "gemini-2.5-flash-tts",
+    "provider": "vertexai",
+    "family": "gemini-2",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "version_id": "default",
+      "open_source_category": null,
+      "launch_stage": "GA",
+      "supported_actions": null,
+      "publisher_model_template": "projects/{project}/locations/{location}/publishers/google/models/gemini-2.5-flash-tts@default"
+    }
+  },
   {
     "id": "gemini-2.5-pro",
     "name": "Gemini 2.5 Pro",
     "provider": "vertexai",
     "family": "gemini-pro",
-    "created_at": "2025-03-20 00:00:00 +0530",
+    "created_at": "2025-03-20 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -23341,18 +60079,24 @@
     "capabilities": [
       "function_calling",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.31
+          "cache_read_input_per_million": 0.125
         }
       }
     },
     "metadata": {
+      "version_id": "default",
+      "open_source_category": null,
+      "launch_stage": "GA",
+      "supported_actions": null,
+      "publisher_model_template": "projects/{project}/locations/{location}/publishers/google/models/gemini-2.5-pro@default",
       "source": "models.dev",
       "provider_id": "google-vertex",
       "open_weights": false,
@@ -23362,7 +60106,12 @@
       "cost": {
         "input": 1.25,
         "output": 10,
-        "cache_read": 0.31
+        "cache_read": 0.125,
+        "context_over_200k": {
+          "input": 2.5,
+          "output": 15,
+          "cache_read": 0.25
+        }
       },
       "limit": {
         "context": 1048576,
@@ -23371,12 +60120,70 @@
       "knowledge": "2025-01"
     }
   },
+  {
+    "id": "gemini-2.5-pro-exp-03-25",
+    "name": "gemini-2.5-pro-exp-03-25",
+    "provider": "vertexai",
+    "family": "gemini-2",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "version_id": "default",
+      "open_source_category": null,
+      "launch_stage": "EXPERIMENTAL",
+      "supported_actions": {
+        "openNotebook": {
+          "references": {
+            "us-central1": {
+              "uri": "https://colab.research.google.com/github/GoogleCloudPlatform/generative-ai/blob/main/gemini/getting-started/intro_gemini_2_5_pro.ipynb"
+            }
+          },
+          "resourceTitle": "Notebook",
+          "resourceUseCase": "Vertex Serving",
+          "resourceDescription": "Intro to Gemini 2.5 Pro."
+        },
+        "openGenerationAiStudio": {
+          "references": {
+            "us-central1": {
+              "uri": "https://console.cloud.google.com/vertex-ai/generative/multimodal/create/text?model=gemini-2.5-pro-exp-03-25"
+            }
+          }
+        },
+        "openNotebooks": {
+          "notebooks": [
+            {
+              "references": {
+                "us-central1": {
+                  "uri": "https://colab.research.google.com/github/GoogleCloudPlatform/generative-ai/blob/main/gemini/getting-started/intro_gemini_2_5_pro.ipynb"
+                }
+              },
+              "resourceTitle": "Notebook",
+              "resourceUseCase": "Vertex Serving",
+              "resourceDescription": "Intro to Gemini 2.5 Pro."
+            }
+          ]
+        }
+      },
+      "publisher_model_template": "projects/{project}/locations/{location}/publishers/google/models/gemini-2.5-pro-exp-03-25@default"
+    }
+  },
   {
     "id": "gemini-2.5-pro-preview-05-06",
     "name": "Gemini 2.5 Pro Preview 05-06",
     "provider": "vertexai",
     "family": "gemini-pro",
-    "created_at": "2025-05-06 00:00:00 +0530",
+    "created_at": "2025-05-06 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -23402,7 +60209,7 @@
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.31
+          "cache_read_input_per_million": 0.31
         }
       }
     },
@@ -23430,7 +60237,7 @@
     "name": "Gemini 2.5 Pro Preview 06-05",
     "provider": "vertexai",
     "family": "gemini-pro",
-    "created_at": "2025-06-05 00:00:00 +0530",
+    "created_at": "2025-06-05 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -23456,7 +60263,7 @@
         "standard": {
           "input_per_million": 1.25,
           "output_per_million": 10,
-          "cached_input_per_million": 0.31
+          "cache_read_input_per_million": 0.31
         }
       }
     },
@@ -23479,12 +60286,38 @@
       "knowledge": "2025-01"
     }
   },
+  {
+    "id": "gemini-2.5-pro-tts",
+    "name": "gemini-2.5-pro-tts",
+    "provider": "vertexai",
+    "family": "gemini-2",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "version_id": "default",
+      "open_source_category": null,
+      "launch_stage": "GA",
+      "supported_actions": null,
+      "publisher_model_template": "projects/{project}/locations/{location}/publishers/google/models/gemini-2.5-pro-tts@default"
+    }
+  },
   {
     "id": "gemini-3-flash-preview",
     "name": "Gemini 3 Flash Preview",
     "provider": "vertexai",
     "family": "gemini-flash",
-    "created_at": "2025-12-17 00:00:00 +0530",
+    "created_at": "2025-12-17 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -23504,18 +60337,24 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 0.5,
           "output_per_million": 3,
-          "cached_input_per_million": 0.05
+          "cache_read_input_per_million": 0.05
         }
       }
     },
     "metadata": {
+      "version_id": "default",
+      "open_source_category": null,
+      "launch_stage": "PUBLIC_PREVIEW",
+      "supported_actions": null,
+      "publisher_model_template": "projects/{project}/locations/{location}/publishers/google/models/gemini-3-flash-preview@default",
       "source": "models.dev",
       "provider_id": "google-vertex",
       "open_weights": false,
@@ -23544,7 +60383,7 @@
     "name": "Gemini 3 Pro Preview",
     "provider": "vertexai",
     "family": "gemini-pro",
-    "created_at": "2025-11-18 00:00:00 +0530",
+    "created_at": "2025-11-18 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -23571,7 +60410,7 @@
         "standard": {
           "input_per_million": 2,
           "output_per_million": 12,
-          "cached_input_per_million": 0.2
+          "cache_read_input_per_million": 0.2
         }
       }
     },
@@ -23600,11 +60439,68 @@
     }
   },
   {
-    "id": "gemini-3.1-pro-preview",
-    "name": "Gemini 3.1 Pro Preview",
+    "id": "gemini-3.1-flash-image-preview",
+    "name": "Gemini 3.1 Flash Image (Preview)",
     "provider": "vertexai",
-    "family": "gemini-pro",
-    "created_at": "2026-02-19 00:00:00 +0530",
+    "family": "gemini-flash",
+    "created_at": "2026-02-26 00:00:00 UTC",
+    "context_window": 131072,
+    "max_output_tokens": 32768,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "pdf"
+      ],
+      "output": [
+        "text",
+        "image"
+      ]
+    },
+    "capabilities": [
+      "reasoning",
+      "vision",
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.25,
+          "output_per_million": 60
+        }
+      }
+    },
+    "metadata": {
+      "version_id": "default",
+      "open_source_category": null,
+      "launch_stage": "PUBLIC_PREVIEW",
+      "supported_actions": null,
+      "publisher_model_template": "projects/{project}/locations/{location}/publishers/google/models/gemini-3.1-flash-image-preview@default",
+      "source": "models.dev",
+      "provider_id": "google",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2026-02-26",
+      "cost": {
+        "input": 0.25,
+        "output": 60
+      },
+      "limit": {
+        "context": 131072,
+        "output": 32768
+      },
+      "knowledge": "2025-01"
+    }
+  },
+  {
+    "id": "gemini-3.1-flash-lite-preview",
+    "name": "Gemini 3.1 Flash Lite Preview",
+    "provider": "vertexai",
+    "family": "gemini-flash-lite",
+    "created_at": "2026-03-03 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -23624,18 +60520,87 @@
       "function_calling",
       "structured_output",
       "reasoning",
-      "vision"
+      "vision",
+      "streaming"
+    ],
+    "pricing": {
+      "text_tokens": {
+        "standard": {
+          "input_per_million": 0.25,
+          "output_per_million": 1.5,
+          "cache_read_input_per_million": 0.025,
+          "cache_write_input_per_million": 1
+        }
+      }
+    },
+    "metadata": {
+      "version_id": "default",
+      "open_source_category": null,
+      "launch_stage": "PUBLIC_PREVIEW",
+      "supported_actions": null,
+      "publisher_model_template": "projects/{project}/locations/{location}/publishers/google/models/gemini-3.1-flash-lite-preview@default",
+      "source": "models.dev",
+      "provider_id": "google-vertex",
+      "open_weights": false,
+      "attachment": true,
+      "temperature": true,
+      "last_updated": "2026-03-03",
+      "cost": {
+        "input": 0.25,
+        "output": 1.5,
+        "cache_read": 0.025,
+        "cache_write": 1
+      },
+      "limit": {
+        "context": 1048576,
+        "output": 65536
+      },
+      "knowledge": "2025-01"
+    }
+  },
+  {
+    "id": "gemini-3.1-pro-preview",
+    "name": "Gemini 3.1 Pro Preview",
+    "provider": "vertexai",
+    "family": "gemini-pro",
+    "created_at": "2026-02-19 00:00:00 UTC",
+    "context_window": 1048576,
+    "max_output_tokens": 65536,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image",
+        "video",
+        "audio",
+        "pdf"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "vision",
+      "streaming"
     ],
     "pricing": {
       "text_tokens": {
         "standard": {
           "input_per_million": 2,
           "output_per_million": 12,
-          "cached_input_per_million": 0.2
+          "cache_read_input_per_million": 0.2
         }
       }
     },
     "metadata": {
+      "version_id": "default",
+      "open_source_category": null,
+      "launch_stage": "PUBLIC_PREVIEW",
+      "supported_actions": null,
+      "publisher_model_template": "projects/{project}/locations/{location}/publishers/google/models/gemini-3.1-pro-preview@default",
       "source": "models.dev",
       "provider_id": "google-vertex",
       "open_weights": false,
@@ -23664,7 +60629,7 @@
     "name": "Gemini 3.1 Pro Preview Custom Tools",
     "provider": "vertexai",
     "family": "gemini-pro",
-    "created_at": "2026-02-19 00:00:00 +0530",
+    "created_at": "2026-02-19 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -23691,7 +60656,7 @@
         "standard": {
           "input_per_million": 2,
           "output_per_million": 12,
-          "cached_input_per_million": 0.2
+          "cache_read_input_per_million": 0.2
         }
       }
     },
@@ -23724,7 +60689,7 @@
     "name": "Gemini Embedding 001",
     "provider": "vertexai",
     "family": "gemini",
-    "created_at": "2025-05-20 00:00:00 +0530",
+    "created_at": "2025-05-20 00:00:00 UTC",
     "context_window": 2048,
     "max_output_tokens": 3072,
     "knowledge_cutoff": null,
@@ -23733,10 +60698,13 @@
         "text"
       ],
       "output": [
-        "text"
+        "embeddings"
       ]
     },
-    "capabilities": [],
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
     "pricing": {
       "text_tokens": {
         "standard": {
@@ -23762,12 +60730,82 @@
       "knowledge": "2025-05"
     }
   },
+  {
+    "id": "gemini-embedding-2",
+    "name": "gemini-embedding-2",
+    "provider": "vertexai",
+    "family": "gemini",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "version_id": "default",
+      "open_source_category": null,
+      "launch_stage": "GA",
+      "supported_actions": null,
+      "publisher_model_template": "projects/{project}/locations/{location}/publishers/google/models/gemini-embedding-2@default"
+    }
+  },
+  {
+    "id": "gemini-exp-1121",
+    "name": "gemini-exp-1121",
+    "provider": "vertexai",
+    "family": "gemini",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "source": "known_models"
+    }
+  },
+  {
+    "id": "gemini-exp-1206",
+    "name": "gemini-exp-1206",
+    "provider": "vertexai",
+    "family": "gemini",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "source": "known_models"
+    }
+  },
   {
     "id": "gemini-flash-latest",
     "name": "Gemini Flash Latest",
     "provider": "vertexai",
     "family": "gemini-flash",
-    "created_at": "2025-09-25 00:00:00 +0530",
+    "created_at": "2025-09-25 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -23793,7 +60831,8 @@
         "standard": {
           "input_per_million": 0.3,
           "output_per_million": 2.5,
-          "cached_input_per_million": 0.075
+          "cache_read_input_per_million": 0.075,
+          "cache_write_input_per_million": 0.383
         }
       }
     },
@@ -23822,7 +60861,7 @@
     "name": "Gemini Flash-Lite Latest",
     "provider": "vertexai",
     "family": "gemini-flash-lite",
-    "created_at": "2025-09-25 00:00:00 +0530",
+    "created_at": "2025-09-25 00:00:00 UTC",
     "context_window": 1048576,
     "max_output_tokens": 65536,
     "knowledge_cutoff": null,
@@ -23848,7 +60887,7 @@
         "standard": {
           "input_per_million": 0.1,
           "output_per_million": 0.4,
-          "cached_input_per_million": 0.025
+          "cache_read_input_per_million": 0.025
         }
       }
     },
@@ -23872,13 +60911,149 @@
     }
   },
   {
-    "id": "meta/llama-3.3-70b-instruct-maas",
-    "name": "Llama 3.3 70B Instruct",
+    "id": "gemini-live-2.5-flash-native-audio",
+    "name": "gemini-live-2.5-flash-native-audio",
     "provider": "vertexai",
-    "family": "llama",
-    "created_at": "2025-04-29 00:00:00 +0530",
-    "context_window": 128000,
-    "max_output_tokens": 8192,
+    "family": "gemini",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "version_id": "default",
+      "open_source_category": null,
+      "launch_stage": "GA",
+      "supported_actions": null,
+      "publisher_model_template": "projects/{project}/locations/{location}/publishers/google/models/gemini-live-2.5-flash-native-audio@default"
+    }
+  },
+  {
+    "id": "gemini-pro",
+    "name": "gemini-pro",
+    "provider": "vertexai",
+    "family": "gemini",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "source": "known_models"
+    }
+  },
+  {
+    "id": "gemini-pro-vision",
+    "name": "gemini-pro-vision",
+    "provider": "vertexai",
+    "family": "gemini",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "source": "known_models"
+    }
+  },
+  {
+    "id": "text-embedding-004",
+    "name": "text-embedding-004",
+    "provider": "vertexai",
+    "family": "text-embedding",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "source": "known_models"
+    }
+  },
+  {
+    "id": "text-embedding-005",
+    "name": "text-embedding-005",
+    "provider": "vertexai",
+    "family": "text-embedding",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "source": "known_models"
+    }
+  },
+  {
+    "id": "text-multilingual-embedding-002",
+    "name": "text-multilingual-embedding-002",
+    "provider": "vertexai",
+    "family": "gemini",
+    "created_at": null,
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [],
+      "output": []
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling"
+    ],
+    "pricing": {},
+    "metadata": {
+      "source": "known_models"
+    }
+  },
+  {
+    "id": "grok-3",
+    "name": "Grok 3",
+    "provider": "xai",
+    "family": "grok",
+    "created_at": "2025-04-04 00:00:00 UTC",
+    "context_window": null,
+    "max_output_tokens": null,
     "knowledge_cutoff": null,
     "modalities": {
       "input": [
@@ -23889,43 +61064,53 @@
       ]
     },
     "capabilities": [
+      "streaming",
       "function_calling",
       "structured_output"
     ],
-    "pricing": {
-      "text_tokens": {
-        "standard": {
-          "input_per_million": 0.72,
-          "output_per_million": 0.72
-        }
-      }
-    },
+    "pricing": {},
     "metadata": {
-      "source": "models.dev",
-      "provider_id": "google-vertex",
-      "open_weights": true,
-      "attachment": false,
-      "temperature": true,
-      "last_updated": "2025-04-29",
-      "cost": {
-        "input": 0.72,
-        "output": 0.72
-      },
-      "limit": {
-        "context": 128000,
-        "output": 8192
-      },
-      "knowledge": "2023-12"
+      "object": "model",
+      "owned_by": "xai"
     }
   },
   {
-    "id": "meta/llama-4-maverick-17b-128e-instruct-maas",
-    "name": "Llama 4 Maverick 17B 128E Instruct",
-    "provider": "vertexai",
-    "family": "llama",
-    "created_at": "2025-04-29 00:00:00 +0530",
-    "context_window": 524288,
-    "max_output_tokens": 8192,
+    "id": "grok-3-mini",
+    "name": "Grok 3 Mini",
+    "provider": "xai",
+    "family": "grok",
+    "created_at": "2025-04-04 00:00:00 UTC",
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "reasoning"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "xai"
+    }
+  },
+  {
+    "id": "grok-4-0709",
+    "name": "Grok 4 0709",
+    "provider": "xai",
+    "family": "grok",
+    "created_at": "2025-07-09 00:00:00 UTC",
+    "context_window": null,
+    "max_output_tokens": null,
     "knowledge_cutoff": null,
     "modalities": {
       "input": [
@@ -23937,287 +61122,148 @@
       ]
     },
     "capabilities": [
-      "function_calling",
-      "structured_output",
-      "vision"
-    ],
-    "pricing": {
-      "text_tokens": {
-        "standard": {
-          "input_per_million": 0.35,
-          "output_per_million": 1.15
-        }
-      }
-    },
-    "metadata": {
-      "source": "models.dev",
-      "provider_id": "google-vertex",
-      "open_weights": true,
-      "attachment": true,
-      "temperature": true,
-      "last_updated": "2025-04-29",
-      "cost": {
-        "input": 0.35,
-        "output": 1.15
-      },
-      "limit": {
-        "context": 524288,
-        "output": 8192
-      },
-      "knowledge": "2024-08"
-    }
-  },
-  {
-    "id": "moonshotai/kimi-k2-thinking-maas",
-    "name": "Kimi K2 Thinking",
-    "provider": "vertexai",
-    "family": "kimi-thinking",
-    "created_at": "2025-11-13 00:00:00 +0530",
-    "context_window": 262144,
-    "max_output_tokens": 262144,
-    "knowledge_cutoff": null,
-    "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
-    },
-    "capabilities": [
-      "function_calling",
-      "structured_output",
-      "reasoning"
-    ],
-    "pricing": {
-      "text_tokens": {
-        "standard": {
-          "input_per_million": 0.6,
-          "output_per_million": 2.5
-        }
-      }
-    },
-    "metadata": {
-      "source": "models.dev",
-      "provider_id": "google-vertex",
-      "open_weights": true,
-      "attachment": false,
-      "temperature": true,
-      "last_updated": "2025-11-13",
-      "interleaved": {
-        "field": "reasoning_content"
-      },
-      "cost": {
-        "input": 0.6,
-        "output": 2.5
-      },
-      "limit": {
-        "context": 262144,
-        "output": 262144
-      },
-      "knowledge": "2024-08"
-    }
-  },
-  {
-    "id": "openai/gpt-oss-120b-maas",
-    "name": "GPT OSS 120B",
-    "provider": "vertexai",
-    "family": "gpt-oss",
-    "created_at": "2025-08-05 00:00:00 +0530",
-    "context_window": 131072,
-    "max_output_tokens": 32768,
-    "knowledge_cutoff": null,
-    "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
-    },
-    "capabilities": [
-      "function_calling",
-      "reasoning"
-    ],
-    "pricing": {
-      "text_tokens": {
-        "standard": {
-          "input_per_million": 0.09,
-          "output_per_million": 0.36
-        }
-      }
-    },
-    "metadata": {
-      "source": "models.dev",
-      "provider_id": "google-vertex",
-      "open_weights": true,
-      "attachment": false,
-      "temperature": true,
-      "last_updated": "2025-08-05",
-      "cost": {
-        "input": 0.09,
-        "output": 0.36
-      },
-      "limit": {
-        "context": 131072,
-        "output": 32768
-      }
-    }
-  },
-  {
-    "id": "openai/gpt-oss-20b-maas",
-    "name": "GPT OSS 20B",
-    "provider": "vertexai",
-    "family": "gpt-oss",
-    "created_at": "2025-08-05 00:00:00 +0530",
-    "context_window": 131072,
-    "max_output_tokens": 32768,
-    "knowledge_cutoff": null,
-    "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
-    },
-    "capabilities": [
-      "function_calling",
-      "reasoning"
-    ],
-    "pricing": {
-      "text_tokens": {
-        "standard": {
-          "input_per_million": 0.07,
-          "output_per_million": 0.25
-        }
-      }
-    },
-    "metadata": {
-      "source": "models.dev",
-      "provider_id": "google-vertex",
-      "open_weights": true,
-      "attachment": false,
-      "temperature": true,
-      "last_updated": "2025-08-05",
-      "cost": {
-        "input": 0.07,
-        "output": 0.25
-      },
-      "limit": {
-        "context": 131072,
-        "output": 32768
-      }
-    }
-  },
-  {
-    "id": "qwen/qwen3-235b-a22b-instruct-2507-maas",
-    "name": "Qwen3 235B A22B Instruct",
-    "provider": "vertexai",
-    "family": "qwen",
-    "created_at": "2025-08-13 00:00:00 +0530",
-    "context_window": 262144,
-    "max_output_tokens": 16384,
-    "knowledge_cutoff": null,
-    "modalities": {
-      "input": [
-        "text"
-      ],
-      "output": [
-        "text"
-      ]
-    },
-    "capabilities": [
-      "function_calling",
-      "structured_output",
-      "reasoning"
-    ],
-    "pricing": {
-      "text_tokens": {
-        "standard": {
-          "input_per_million": 0.22,
-          "output_per_million": 0.88
-        }
-      }
-    },
-    "metadata": {
-      "source": "models.dev",
-      "provider_id": "google-vertex",
-      "open_weights": true,
-      "attachment": false,
-      "temperature": true,
-      "last_updated": "2025-08-13",
-      "cost": {
-        "input": 0.22,
-        "output": 0.88
-      },
-      "limit": {
-        "context": 262144,
-        "output": 16384
-      }
-    }
-  },
-  {
-    "id": "zai-org/glm-4.7-maas",
-    "name": "GLM-4.7",
-    "provider": "vertexai",
-    "family": "glm",
-    "created_at": "2026-01-06 00:00:00 +0530",
-    "context_window": 200000,
-    "max_output_tokens": 128000,
-    "knowledge_cutoff": null,
-    "modalities": {
-      "input": [
-        "text",
-        "pdf"
-      ],
-      "output": [
-        "text"
-      ]
-    },
-    "capabilities": [
+      "streaming",
       "function_calling",
       "structured_output",
       "reasoning",
       "vision"
     ],
-    "pricing": {
-      "text_tokens": {
-        "standard": {
-          "input_per_million": 0.6,
-          "output_per_million": 2.2
-        }
-      }
-    },
+    "pricing": {},
     "metadata": {
-      "source": "models.dev",
-      "provider_id": "google-vertex",
-      "open_weights": true,
-      "attachment": false,
-      "temperature": true,
-      "last_updated": "2026-01-06",
-      "interleaved": {
-        "field": "reasoning_content"
-      },
-      "cost": {
-        "input": 0.6,
-        "output": 2.2
-      },
-      "limit": {
-        "context": 200000,
-        "output": 128000
-      },
-      "knowledge": "2025-04"
+      "object": "model",
+      "owned_by": "xai"
     }
   },
   {
-    "id": "zai-org/glm-5-maas",
-    "name": "GLM-5",
-    "provider": "vertexai",
-    "family": "glm",
-    "created_at": "2026-02-11 00:00:00 +0530",
-    "context_window": 202752,
-    "max_output_tokens": 131072,
+    "id": "grok-4-1-fast-non-reasoning",
+    "name": "Grok 4 1 Fast Non Reasoning",
+    "provider": "xai",
+    "family": "grok",
+    "created_at": "2025-11-19 00:00:00 UTC",
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "xai"
+    }
+  },
+  {
+    "id": "grok-4-1-fast-reasoning",
+    "name": "Grok 4 1 Fast Reasoning",
+    "provider": "xai",
+    "family": "grok",
+    "created_at": "2025-11-19 00:00:00 UTC",
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "xai"
+    }
+  },
+  {
+    "id": "grok-4-fast-non-reasoning",
+    "name": "Grok 4 Fast Non Reasoning",
+    "provider": "xai",
+    "family": "grok",
+    "created_at": "2025-09-04 00:00:00 UTC",
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "vision"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "xai"
+    }
+  },
+  {
+    "id": "grok-4-fast-reasoning",
+    "name": "Grok 4 Fast Reasoning",
+    "provider": "xai",
+    "family": "grok",
+    "created_at": "2025-09-04 00:00:00 UTC",
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text",
+        "image"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
+      "reasoning",
+      "vision"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "xai"
+    }
+  },
+  {
+    "id": "grok-4.20-0309-non-reasoning",
+    "name": "Grok 4.20 0309 Non Reasoning",
+    "provider": "xai",
+    "family": "grok",
+    "created_at": "2026-03-09 00:00:00 UTC",
+    "context_window": null,
+    "max_output_tokens": null,
     "knowledge_cutoff": null,
     "modalities": {
       "input": [
@@ -24228,37 +61274,239 @@
       ]
     },
     "capabilities": [
+      "streaming",
       "function_calling",
+      "structured_output"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "xai"
+    }
+  },
+  {
+    "id": "grok-4.20-0309-reasoning",
+    "name": "Grok 4.20 0309 Reasoning",
+    "provider": "xai",
+    "family": "grok",
+    "created_at": "2026-03-09 00:00:00 UTC",
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "xai"
+    }
+  },
+  {
+    "id": "grok-4.20-multi-agent-0309",
+    "name": "Grok 4.20 Multi Agent 0309",
+    "provider": "xai",
+    "family": "grok",
+    "created_at": "2026-03-09 00:00:00 UTC",
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "xai"
+    }
+  },
+  {
+    "id": "grok-4.3",
+    "name": "Grok 4.3",
+    "provider": "xai",
+    "family": "grok",
+    "created_at": "2026-04-17 00:00:00 UTC",
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "xai"
+    }
+  },
+  {
+    "id": "grok-code-fast-1",
+    "name": "Grok Code Fast 1",
+    "provider": "xai",
+    "family": "grok",
+    "created_at": "2025-08-24 00:00:00 UTC",
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output",
       "reasoning"
     ],
-    "pricing": {
-      "text_tokens": {
-        "standard": {
-          "input_per_million": 1,
-          "output_per_million": 3.2,
-          "cached_input_per_million": 0.1
-        }
-      }
-    },
+    "pricing": {},
     "metadata": {
-      "source": "models.dev",
-      "provider_id": "google-vertex",
-      "open_weights": true,
-      "attachment": false,
-      "temperature": true,
-      "last_updated": "2026-02-11",
-      "interleaved": {
-        "field": "reasoning_content"
-      },
-      "cost": {
-        "input": 1,
-        "output": 3.2,
-        "cache_read": 0.1
-      },
-      "limit": {
-        "context": 202752,
-        "output": 131072
-      }
+      "object": "model",
+      "owned_by": "xai"
+    }
+  },
+  {
+    "id": "grok-imagine-image",
+    "name": "Grok Imagine Image",
+    "provider": "xai",
+    "family": "grok",
+    "created_at": "2026-01-28 00:00:00 UTC",
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "xai"
+    }
+  },
+  {
+    "id": "grok-imagine-image-pro",
+    "name": "Grok Imagine Image Pro",
+    "provider": "xai",
+    "family": "grok",
+    "created_at": "2026-01-28 00:00:00 UTC",
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "xai"
+    }
+  },
+  {
+    "id": "grok-imagine-image-quality",
+    "name": "Grok Imagine Image Quality",
+    "provider": "xai",
+    "family": "grok",
+    "created_at": "2026-04-03 00:00:00 UTC",
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "xai"
+    }
+  },
+  {
+    "id": "grok-imagine-video",
+    "name": "Grok Imagine Video",
+    "provider": "xai",
+    "family": "grok",
+    "created_at": "2026-01-28 00:00:00 UTC",
+    "context_window": null,
+    "max_output_tokens": null,
+    "knowledge_cutoff": null,
+    "modalities": {
+      "input": [
+        "text"
+      ],
+      "output": [
+        "text"
+      ]
+    },
+    "capabilities": [
+      "streaming",
+      "function_calling",
+      "structured_output"
+    ],
+    "pricing": {},
+    "metadata": {
+      "object": "model",
+      "owned_by": "xai"
     }
   }
 ]
\ No newline at end of file

From cc612e755b5bda1bfe67d478c62798a19f3147b5 Mon Sep 17 00:00:00 2001
From: Shivam Mishra <scm.mymail@gmail.com>
Date: Fri, 8 May 2026 21:18:50 +0530
Subject: [PATCH 08/20] fix: SafeFetch dependency loading (#14408)

The SafeFetch spec suite was failing in CI with `NameError:
uninitialized constant SafeFetch::Fetcher` across every example that
exercised `SafeFetch.fetch`. From a product perspective, this made the
external-file fetch path look unreliable even though the failure
happened before any network validation, SSRF protection, content-type
checks, or tempfile handling could run.

The symptom pointed to a load-order issue rather than an actual fetch
behavior regression. `SafeFetch.fetch` referenced `Fetcher` from the
top-level module, but that nested class was not guaranteed to be loaded
in every test execution path before the method was invoked.

This change keeps the existing SafeFetch split between the public API
and the implementation classes, but makes the public entry point
responsible for loading the implementation it needs before use. That is
intentionally smaller than folding all of the fetcher logic into
`lib/safe_fetch.rb`; the separate files still keep the request option
parsing and streaming implementation readable, while the public API no
longer depends on Rails or the test runner having loaded nested
constants in a particular order.

The file also now uses a single `SafeFetch` module declaration. That
removes the awkward reopen pattern and makes the dependency boundary
easier to see: constants and errors are defined first, then the public
`fetch` method loads and delegates to the implementation classes.
---
 lib/safe_fetch.rb | 7 +------
 1 file changed, 1 insertion(+), 6 deletions(-)

diff --git a/lib/safe_fetch.rb b/lib/safe_fetch.rb
index d664dcf6a..7f89c03c4 100644
--- a/lib/safe_fetch.rb
+++ b/lib/safe_fetch.rb
@@ -22,16 +22,11 @@ module SafeFetch
   class FileTooLargeError < Error; end
   class UnsupportedContentTypeError < Error; end
   class UnsupportedMethodError < Error; end
-end
 
-require_relative 'safe_fetch/request_options'
-require_relative 'safe_fetch/fetcher'
-
-module SafeFetch
   def self.fetch(url, **, &)
     raise ArgumentError, 'block required' unless block_given?
 
-    Fetcher.new(RequestOptions.new(url: url, **)).fetch(&)
+    SafeFetch::Fetcher.new(SafeFetch::RequestOptions.new(url: url, **)).fetch(&)
   rescue SsrfFilter::InvalidUriScheme, URI::InvalidURIError => e
     raise InvalidUrlError, e.message
   rescue SsrfFilter::Error, Resolv::ResolvError => e

From bc768bf04f4199bebf6d0e30ed226e02405deeb3 Mon Sep 17 00:00:00 2001
From: Shivam Mishra <scm.mymail@gmail.com>
Date: Mon, 11 May 2026 10:58:23 +0530
Subject: [PATCH 09/20] chore: verbosely log errors for leadsquare activity
 failure (#14407)

---
 .../crm/leadsquared/processor_service.rb         | 16 ++++++++++++----
 .../crm/leadsquared/processor_service_spec.rb    |  2 +-
 2 files changed, 13 insertions(+), 5 deletions(-)

diff --git a/app/services/crm/leadsquared/processor_service.rb b/app/services/crm/leadsquared/processor_service.rb
index ef33718f2..9ffa3d12c 100644
--- a/app/services/crm/leadsquared/processor_service.rb
+++ b/app/services/crm/leadsquared/processor_service.rb
@@ -89,11 +89,19 @@ class Crm::Leadsquared::ProcessorService < Crm::BaseProcessorService
     metadata[metadata_key] = activity_id
     store_conversation_metadata(conversation, metadata)
   rescue Crm::Leadsquared::Api::BaseClient::ApiError => e
-    ChatwootExceptionTracker.new(e, account: @account).capture_exception
-    Rails.logger.error "LeadSquared API error in #{activity_type} activity: #{e.message}"
+    log_activity_error(e, activity_type, conversation, payload: { lead_id: lead_id, activity_code: activity_code, activity_note: activity_note })
   rescue StandardError => e
-    ChatwootExceptionTracker.new(e, account: @account).capture_exception
-    Rails.logger.error "Error creating #{activity_type} activity in LeadSquared: #{e.message}"
+    log_activity_error(e, activity_type, conversation)
+  end
+
+  def log_activity_error(error, activity_type, conversation, payload: nil)
+    ChatwootExceptionTracker.new(error, account: @account).capture_exception
+    context = "account_id=#{conversation.account_id}, conversation_display_id=#{conversation.display_id}"
+    if payload
+      context += ", http_status=#{error.code}, prospect_id=#{payload[:lead_id]}, " \
+                 "activity_event=#{payload[:activity_code]}, note_bytes=#{payload[:activity_note].to_s.bytesize}"
+    end
+    Rails.logger.error("LeadSquared #{activity_type} activity failed: #{error.message} (#{context})")
   end
 
   def get_activity_code(key)
diff --git a/spec/services/crm/leadsquared/processor_service_spec.rb b/spec/services/crm/leadsquared/processor_service_spec.rb
index 7008eb064..7b99721c5 100644
--- a/spec/services/crm/leadsquared/processor_service_spec.rb
+++ b/spec/services/crm/leadsquared/processor_service_spec.rb
@@ -157,7 +157,7 @@ RSpec.describe Crm::Leadsquared::ProcessorService do
 
         it 'logs the error' do
           service.handle_conversation_created(conversation)
-          expect(Rails.logger).to have_received(:error).with(/Error creating conversation activity/)
+          expect(Rails.logger).to have_received(:error).with(/LeadSquared conversation activity failed/)
         end
       end
     end

From 2e13f69fdf61805fbd6f3cb2cfe7813fed744648 Mon Sep 17 00:00:00 2001
From: Shivam Mishra <scm.mymail@gmail.com>
Date: Mon, 11 May 2026 16:09:45 +0530
Subject: [PATCH 10/20] chore: log errors from context.dev (#14310)

This PR updates the way we log errors and results from context.dev to
have better visibility on the enrichment process for onboarding
---
 app/jobs/account/branding_enrichment_job.rb                  | 5 ++++-
 .../app/services/enterprise/website_branding_service.rb      | 5 ++++-
 2 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/app/jobs/account/branding_enrichment_job.rb b/app/jobs/account/branding_enrichment_job.rb
index 2898604ca..3deb81821 100644
--- a/app/jobs/account/branding_enrichment_job.rb
+++ b/app/jobs/account/branding_enrichment_job.rb
@@ -3,7 +3,10 @@ class Account::BrandingEnrichmentJob < ApplicationJob
 
   def perform(account_id, email)
     result = WebsiteBrandingService.new(email).perform
-    return if result.blank?
+    if result.blank?
+      Rails.logger.info "[BrandingEnrichment] Enrichment failed for account=#{account_id} email=#{email}"
+      return
+    end
 
     account = Account.find(account_id)
     account.name = result[:title] if result[:title].present?
diff --git a/enterprise/app/services/enterprise/website_branding_service.rb b/enterprise/app/services/enterprise/website_branding_service.rb
index a1925e80d..553317c43 100644
--- a/enterprise/app/services/enterprise/website_branding_service.rb
+++ b/enterprise/app/services/enterprise/website_branding_service.rb
@@ -34,7 +34,10 @@ module Enterprise::WebsiteBrandingService
 
   def process_response(response)
     @http_status = response.code
-    raise "API Error: #{response.message} (Status: #{response.code})" unless response.success?
+    unless response.success?
+      Rails.logger.warn "[WebsiteBranding] Context.dev returned #{response.code}: #{response.parsed_response}"
+      return nil
+    end
 
     brand = response.parsed_response&.dig('brand')
     return nil if brand.blank?

From 34892987268d22230e29baf5d60cc4e2961a65a4 Mon Sep 17 00:00:00 2001
From: Shivam Mishra <scm.mymail@gmail.com>
Date: Mon, 11 May 2026 16:10:48 +0530
Subject: [PATCH 11/20] feat: add `WidgetCreationService` for onboarding web
 widget setup (#14314)

When a new account finishes onboarding we want to land them on a
dashboard with a working web widget already configured, branded, named,
and assigned to them, instead of an empty inbox list. This PR adds the
services that produce that widget. **No user-visible change yet:** the
services are dormant until the trigger and background job are wired up
in the follow-up PR.

## Context

Milestone 1 added `Account::BrandingEnrichmentJob`, which calls
context.dev during signup and stores brand data on
`account.custom_attributes['brand_info']`, plus the new onboarding form
that captures `domain`, `name`, `industry`, etc. Milestone 2 starts
using that data, and the first thing we want is a web widget
materialized automatically. Splitting the service layer from the
orchestration plumbing (Redis key, `onboarding_step` extension,
controller wiring, ActionCable) keeps this diff focused and lets the
LLM/widget logic merge independently.

## How to test

Run against an existing account that already has `brand_info` populated.

```ruby
account = Account.find(<account_id>)
user    = account.administrators.first
inbox   = WidgetCreationService.new(account, user).perform

inbox.channel.widget_color     # color from brand_info, or '#1f93ff'
inbox.channel.welcome_title    # brand_info[:title], or account.name
inbox.channel.welcome_tagline  # LLM tagline (Enterprise + system key set),
                               # else brand_info[:slogan]/[:description]/nil
inbox.inbox_members.pluck(:user_id)
```

Toggle `InstallationConfig['CAPTAIN_OPEN_AI_API_KEY']` to flip between
LLM and brand-text tagline paths. To verify failure isolation, raise
inside `Captain::Llm::WidgetTaglineService#perform` and confirm widget
creation still succeeds with the fallback tagline.
---
 .../onboarding/web_widget_creation_service.rb | 75 ++++++++++++++++++
 .../captain/llm/widget_tagline_schema.rb      |  5 ++
 .../captain/llm/widget_tagline_service.rb     | 78 +++++++++++++++++++
 .../onboarding/web_widget_creation_service.rb | 11 +++
 .../enterprise/captain/base_task_service.rb   |  4 +-
 lib/captain/base_task_service.rb              | 10 +++
 .../lib/captain/base_task_service_spec.rb     | 32 ++++++++
 .../web_widget_creation_service_spec.rb       | 59 ++++++++++++++
 8 files changed, 272 insertions(+), 2 deletions(-)
 create mode 100644 app/services/onboarding/web_widget_creation_service.rb
 create mode 100644 enterprise/app/services/captain/llm/widget_tagline_schema.rb
 create mode 100644 enterprise/app/services/captain/llm/widget_tagline_service.rb
 create mode 100644 enterprise/app/services/enterprise/onboarding/web_widget_creation_service.rb
 create mode 100644 spec/enterprise/services/enterprise/onboarding/web_widget_creation_service_spec.rb

diff --git a/app/services/onboarding/web_widget_creation_service.rb b/app/services/onboarding/web_widget_creation_service.rb
new file mode 100644
index 000000000..363643f05
--- /dev/null
+++ b/app/services/onboarding/web_widget_creation_service.rb
@@ -0,0 +1,75 @@
+class Onboarding::WebWidgetCreationService
+  DEFAULT_WIDGET_COLOR = '#1f93ff'.freeze
+  # context.dev descriptions and LLM completions are unbounded; bound the
+  # stored tagline so a long string doesn't render as a wall of text in the
+  # widget UI (and so backends that enforce varchar limits don't raise).
+  WELCOME_TAGLINE_MAX_LENGTH = 255
+
+  def initialize(account, user)
+    @account = account
+    @user = user
+  end
+
+  def perform
+    existing = existing_web_widget_inbox
+    if existing
+      Rails.logger.info "[WidgetCreation] Reusing existing web widget inbox #{existing.id} for account #{@account.id}"
+      return existing
+    end
+
+    if website_url.blank?
+      Rails.logger.info "[WidgetCreation] Skipping for account #{@account.id}: no website_url available"
+      return nil
+    end
+
+    attrs = channel_attributes
+
+    ActiveRecord::Base.transaction do
+      channel = @account.web_widgets.create!(attrs)
+      inbox = @account.inboxes.create!(name: @account.name, channel: channel)
+      InboxMember.find_or_create_by!(inbox: inbox, user: @user)
+      inbox
+    end
+  rescue StandardError => e
+    Rails.logger.error "[WidgetCreation] #{e.message}"
+    nil
+  end
+
+  private
+
+  def existing_web_widget_inbox
+    @account.inboxes.find_by(channel_type: 'Channel::WebWidget')
+  end
+
+  def channel_attributes
+    {
+      website_url: website_url,
+      widget_color: widget_color,
+      welcome_title: welcome_title,
+      welcome_tagline: welcome_tagline_text&.truncate(WELCOME_TAGLINE_MAX_LENGTH)
+    }
+  end
+
+  def brand_info
+    @brand_info ||= (@account.custom_attributes['brand_info'] || {}).deep_symbolize_keys
+  end
+
+  def website_url
+    @account.domain.presence || brand_info[:domain].presence
+  end
+
+  def widget_color
+    hex = brand_info[:colors]&.first&.dig(:hex)
+    hex.to_s.match?(/\A#\h{6}\z/) ? hex : DEFAULT_WIDGET_COLOR
+  end
+
+  def welcome_title
+    brand_info[:title].presence || @account.name
+  end
+
+  def welcome_tagline_text
+    brand_info[:slogan].presence || brand_info[:description].presence
+  end
+end
+
+Onboarding::WebWidgetCreationService.prepend_mod_with('Onboarding::WebWidgetCreationService')
diff --git a/enterprise/app/services/captain/llm/widget_tagline_schema.rb b/enterprise/app/services/captain/llm/widget_tagline_schema.rb
new file mode 100644
index 000000000..cfb1bc4c5
--- /dev/null
+++ b/enterprise/app/services/captain/llm/widget_tagline_schema.rb
@@ -0,0 +1,5 @@
+class Captain::Llm::WidgetTaglineSchema < RubyLLM::Schema
+  string :tagline,
+         description: 'Short marketing tagline for a customer-support chat widget. Plain text, no quotes, no emoji, no trailing punctuation.',
+         max_length: 60
+end
diff --git a/enterprise/app/services/captain/llm/widget_tagline_service.rb b/enterprise/app/services/captain/llm/widget_tagline_service.rb
new file mode 100644
index 000000000..230c54165
--- /dev/null
+++ b/enterprise/app/services/captain/llm/widget_tagline_service.rb
@@ -0,0 +1,78 @@
+class Captain::Llm::WidgetTaglineService < Captain::BaseTaskService
+  RESPONSE_SCHEMA = Captain::Llm::WidgetTaglineSchema
+
+  pattr_initialize [:account!]
+
+  def perform
+    response = make_api_call(model: tagline_model, messages: messages, schema: RESPONSE_SCHEMA)
+    return response if response[:error]
+
+    response.merge(message: extract_tagline(response[:message]))
+  end
+
+  private
+
+  def extract_tagline(message)
+    tagline = message.is_a?(Hash) ? (message['tagline'] || message[:tagline]) : message
+    tagline.to_s.strip
+  end
+
+  def messages
+    [
+      { role: 'system', content: system_prompt },
+      { role: 'user', content: user_prompt }
+    ]
+  end
+
+  def system_prompt
+    <<~PROMPT
+      You write a short marketing tagline for a company's customer-support chat widget.
+      Use the provided company context to make the tagline specific and on-brand.
+    PROMPT
+  end
+
+  def user_prompt
+    parts = [
+      "Company: #{account.name}",
+      ("Title: #{brand_info[:title]}" if brand_info[:title].present?),
+      ("Description: #{brand_info[:description]}" if brand_info[:description].present?),
+      ("Slogan: #{brand_info[:slogan]}" if brand_info[:slogan].present?),
+      ("Industries: #{industries_text}" if industries_text.present?)
+    ].compact
+    parts.join("\n")
+  end
+
+  def brand_info
+    @brand_info ||= (account.custom_attributes['brand_info'] || {}).deep_symbolize_keys
+  end
+
+  def industries_text
+    Array(brand_info[:industries]).filter_map { |i| i.is_a?(Hash) ? i[:industry] : i }.join(', ').presence
+  end
+
+  def event_name
+    'widget_tagline'
+  end
+
+  def llm_credential
+    @llm_credential ||= system_llm_credential
+  end
+
+  def captain_tasks_enabled?
+    true
+  end
+
+  # Tagline generation runs on the operator's OpenAI key during onboarding;
+  # the customer should not have captain_responses quota deducted for it.
+  def counts_toward_usage?
+    false
+  end
+
+  def tagline_model
+    @tagline_model ||= InstallationConfig.find_by(name: 'CAPTAIN_OPEN_AI_MODEL')&.value.presence || GPT_MODEL
+  end
+
+  def build_follow_up_context?
+    false
+  end
+end
diff --git a/enterprise/app/services/enterprise/onboarding/web_widget_creation_service.rb b/enterprise/app/services/enterprise/onboarding/web_widget_creation_service.rb
new file mode 100644
index 000000000..b108dd5e7
--- /dev/null
+++ b/enterprise/app/services/enterprise/onboarding/web_widget_creation_service.rb
@@ -0,0 +1,11 @@
+module Enterprise::Onboarding::WebWidgetCreationService
+  private
+
+  def welcome_tagline_text
+    response = Captain::Llm::WidgetTaglineService.new(account: @account).perform
+    response&.dig(:message).to_s.strip.presence || super
+  rescue StandardError => e
+    Rails.logger.error "[WidgetCreation] LLM tagline failed: #{e.message}"
+    super
+  end
+end
diff --git a/enterprise/lib/enterprise/captain/base_task_service.rb b/enterprise/lib/enterprise/captain/base_task_service.rb
index 9845359f5..427aeb06d 100644
--- a/enterprise/lib/enterprise/captain/base_task_service.rb
+++ b/enterprise/lib/enterprise/captain/base_task_service.rb
@@ -1,6 +1,6 @@
 module Enterprise::Captain::BaseTaskService
   def perform
-    return { error: I18n.t('captain.copilot_limit'), error_code: 429 } unless responses_available?
+    return { error: I18n.t('captain.copilot_limit'), error_code: 429 } if counts_toward_usage? && !responses_available?
 
     unless captain_tasks_enabled?
       return { error: I18n.t('captain.upgrade') } if ChatwootApp.chatwoot_cloud?
@@ -9,7 +9,7 @@ module Enterprise::Captain::BaseTaskService
     end
 
     result = super
-    increment_usage if successful_result?(result)
+    increment_usage if counts_toward_usage? && successful_result?(result)
     result
   end
 
diff --git a/lib/captain/base_task_service.rb b/lib/captain/base_task_service.rb
index 123377ea0..a043d38e2 100644
--- a/lib/captain/base_task_service.rb
+++ b/lib/captain/base_task_service.rb
@@ -149,6 +149,16 @@ class Captain::BaseTaskService
     account.feature_enabled?('captain_tasks')
   end
 
+  # Extension point consulted by the Enterprise quota wrapper. Subclasses
+  # whose calls run on the operator's key (e.g. internal/onboarding tasks)
+  # should override this to return false. When false, the wrapper neither
+  # blocks the call on an exhausted captain_responses quota nor decrements
+  # it on success — the call participates in the quota system in neither
+  # direction.
+  def counts_toward_usage?
+    true
+  end
+
   def api_key_configured?
     llm_credential.present?
   end
diff --git a/spec/enterprise/lib/captain/base_task_service_spec.rb b/spec/enterprise/lib/captain/base_task_service_spec.rb
index a018a7c84..b3dc473eb 100644
--- a/spec/enterprise/lib/captain/base_task_service_spec.rb
+++ b/spec/enterprise/lib/captain/base_task_service_spec.rb
@@ -165,5 +165,37 @@ RSpec.describe Captain::BaseTaskService, type: :model do
         service.perform
       end
     end
+
+    context 'when subclass opts out via counts_toward_usage?' do
+      let(:test_service_class) do
+        result = perform_result
+        klass = Class.new(described_class) do
+          define_method(:perform) { result }
+          define_method(:event_name) { 'test_event' }
+          define_method(:counts_toward_usage?) { false }
+        end
+        klass.prepend(Enterprise::Captain::BaseTaskService)
+        klass
+      end
+
+      it 'does not increment usage even on a successful result' do
+        expect(account).not_to receive(:increment_response_usage)
+        service.perform
+      end
+
+      context 'when the captain_responses quota is exhausted on Cloud' do
+        before do
+          allow(ChatwootApp).to receive(:chatwoot_cloud?).and_return(true)
+          allow(account).to receive(:usage_limits).and_return({
+                                                                captain: { responses: { current_available: 0 } }
+                                                              })
+        end
+
+        it 'bypasses the 429 gate and returns the underlying result' do
+          result = service.perform
+          expect(result).to eq(perform_result)
+        end
+      end
+    end
   end
 end
diff --git a/spec/enterprise/services/enterprise/onboarding/web_widget_creation_service_spec.rb b/spec/enterprise/services/enterprise/onboarding/web_widget_creation_service_spec.rb
new file mode 100644
index 000000000..90272f5ef
--- /dev/null
+++ b/spec/enterprise/services/enterprise/onboarding/web_widget_creation_service_spec.rb
@@ -0,0 +1,59 @@
+require 'rails_helper'
+
+# Simulate the prepend_mod_with overlay for testing.
+test_klass = Class.new(Onboarding::WebWidgetCreationService) do
+  prepend Enterprise::Onboarding::WebWidgetCreationService
+end
+
+RSpec.describe Enterprise::Onboarding::WebWidgetCreationService do
+  let(:account) do
+    create(:account, name: 'Acme Inc', domain: 'acme.com', custom_attributes: {
+             'brand_info' => { 'slogan' => 'Fallback slogan', 'description' => 'Fallback description' }
+           })
+  end
+  let(:user) { create(:user) }
+  let(:service) { test_klass.new(account, user) }
+
+  before { create(:account_user, account: account, user: user, role: :administrator) }
+
+  describe '#welcome_tagline_text via #perform' do
+    let(:llm_double) { instance_double(Captain::Llm::WidgetTaglineService) }
+
+    before do
+      allow(Captain::Llm::WidgetTaglineService).to receive(:new).and_return(llm_double)
+    end
+
+    context 'when the LLM returns a tagline' do
+      before { allow(llm_double).to receive(:perform).and_return(message: '  LLM tagline  ') }
+
+      it 'uses the (stripped) LLM-generated tagline' do
+        expect(service.perform.channel.welcome_tagline).to eq('LLM tagline')
+      end
+    end
+
+    context 'when the LLM returns a blank message' do
+      before { allow(llm_double).to receive(:perform).and_return(message: '') }
+
+      it 'falls back to brand_info text' do
+        expect(service.perform.channel.welcome_tagline).to eq('Fallback slogan')
+      end
+    end
+
+    context 'when the LLM returns an error response' do
+      before { allow(llm_double).to receive(:perform).and_return(error: 'LLM timeout', error_code: 500) }
+
+      it 'falls back to brand_info text' do
+        expect(service.perform.channel.welcome_tagline).to eq('Fallback slogan')
+      end
+    end
+
+    context 'when the LLM raises an exception' do
+      before { allow(llm_double).to receive(:perform).and_raise(StandardError, 'boom') }
+
+      it 'still creates the widget with brand_info fallback (no transaction rollback)' do
+        expect { service.perform }.to change(Channel::WebWidget, :count).by(1)
+        expect(service.perform.channel.welcome_tagline).to eq('Fallback slogan')
+      end
+    end
+  end
+end

From f6be0d80efbf373e166f7f5b4211016c659d7a78 Mon Sep 17 00:00:00 2001
From: Aakash Bakhle <48802744+aakashb95@users.noreply.github.com>
Date: Mon, 11 May 2026 20:13:29 +0530
Subject: [PATCH 12/20] feat: UI changes for document auto sync [AI-153]
 (#14258)

# Pull Request Template

## Description

FE code for document sync

Adds:
- UI to show counts (stats) of available web pages, stale and synced
documents and last synced at
- Bulk action and manual ways to sync web documents
- index to stats related columns

## Type of change

Please delete options that are not relevant.

- [x] New feature (non-breaking change which adds functionality)

## How Has This Been Tested?

Please describe the tests that you ran to verify your changes. Provide
instructions so we can reproduce. Please also list any relevant details
for your test configuration.

https://linear.app/chatwoot/issue/AI-153/fe-document-auto-sync

Documents dashboard:
<img width="2160" height="986" alt="CleanShot 2026-05-11 at 17 57 09@2x"
src="https://github.com/user-attachments/assets/6d934764-964c-4656-b005-1b4f0329e553"
/>

Filters:
<img width="1138" height="564" alt="CleanShot 2026-05-11 at 17 58 13@2x"
src="https://github.com/user-attachments/assets/cee780e6-eb8f-4aed-8cc5-b674244a821b"
/>

Needs update:
<img width="2222" height="966" alt="CleanShot 2026-05-11 at 17 57 53@2x"
src="https://github.com/user-attachments/assets/70c85ddd-7eb1-4328-ba14-7929e67e7b36"
/>

pdfs:
<img width="2180" height="558" alt="CleanShot 2026-05-11 at 17 58 30@2x"
src="https://github.com/user-attachments/assets/975b5c9f-bd1c-4979-9870-8f926d7f6e11"
/>

bulk actions:
<img width="2244" height="992" alt="CleanShot 2026-05-11 at 17 58 57@2x"
src="https://github.com/user-attachments/assets/bdb3c63f-d2de-41dc-a6d5-8821d3303be0"
/>

single url sync:
<img width="2264" height="722" alt="CleanShot 2026-05-11 at 17 59 19@2x"
src="https://github.com/user-attachments/assets/7d7323a5-0fcb-4be9-8635-55e56964999b"
/>



## Checklist:

- [x] My code follows the style guidelines of this project
- [x] I have performed a self-review of my code
- [x] I have commented on my code, particularly in hard-to-understand
areas
- [ ] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
- [x] I have added tests that prove my fix is effective or that my
feature works
- [x] New and existing unit tests pass locally with my changes
- [x] Any dependent changes have been merged and published in downstream
modules

---------

Co-authored-by: Sivin Varghese <64252451+iamsivin@users.noreply.github.com>
Co-authored-by: iamsivin <iamsivin@gmail.com>
Co-authored-by: Muhsin Keloth <muhsinkeramam@gmail.com>
Co-authored-by: Sony Mathew <sony@chatwoot.com>
Co-authored-by: Vishnu Narayanan <iamwishnu@gmail.com>
---
 .../dashboard/api/captain/document.js         |  11 +-
 .../captain/assistant/DocumentBulkActions.vue | 113 +++++++
 .../captain/assistant/DocumentCard.vue        | 107 ++++++-
 .../captain/assistant/DocumentFilter.vue      |  76 +++++
 .../captain/assistant/DocumentFiltersBar.vue  | 165 +++++++++++
 .../captain/assistant/DocumentSyncStatus.vue  | 153 ++++++++++
 .../document/CreateDocumentDialog.vue         |   3 +-
 app/javascript/dashboard/featureFlags.js      |   1 +
 .../i18n/locale/en/integrations.json          |  50 ++++
 .../dashboard/captain/documents/Index.vue     | 276 ++++++++++++++----
 .../dashboard/store/captain/bulkActions.js    |  13 +
 .../dashboard/store/captain/document.js       |  40 +++
 .../shared/helpers/documentHelper.js          |  44 ++-
 .../helpers/specs/documentHelper.spec.js      |  48 ++-
 ...d_sync_stats_index_to_captain_documents.rb |  12 +
 db/schema.rb                                  |   3 +-
 .../captain/bulk_actions_controller.rb        |   7 +-
 .../accounts/captain/documents_controller.rb  |  67 ++++-
 .../captain/documents/perform_sync_job.rb     |  26 +-
 .../captain/documents/schedule_syncs_job.rb   |  11 +-
 enterprise/app/models/captain/document.rb     |  11 +
 .../app/policies/captain/assistant_policy.rb  |   4 +
 .../captain/documents/sync_service.rb         |  12 +-
 .../captain/documents/index.json.jbuilder     |   1 +
 .../v1/models/captain/_document.json.jbuilder |   2 +
 25 files changed, 1155 insertions(+), 101 deletions(-)
 create mode 100644 app/javascript/dashboard/components-next/captain/assistant/DocumentBulkActions.vue
 create mode 100644 app/javascript/dashboard/components-next/captain/assistant/DocumentFilter.vue
 create mode 100644 app/javascript/dashboard/components-next/captain/assistant/DocumentFiltersBar.vue
 create mode 100644 app/javascript/dashboard/components-next/captain/assistant/DocumentSyncStatus.vue
 create mode 100644 db/migrate/20260429043000_add_sync_stats_index_to_captain_documents.rb

diff --git a/app/javascript/dashboard/api/captain/document.js b/app/javascript/dashboard/api/captain/document.js
index dc22b0c32..e23a8c460 100644
--- a/app/javascript/dashboard/api/captain/document.js
+++ b/app/javascript/dashboard/api/captain/document.js
@@ -6,15 +6,22 @@ class CaptainDocument extends ApiClient {
     super('captain/documents', { accountScoped: true });
   }
 
-  get({ page = 1, searchKey, assistantId } = {}) {
+  get({ page = 1, searchKey, assistantId, filter, source, sort } = {}) {
     return axios.get(this.url, {
       params: {
         page,
-        searchKey,
+        search_key: searchKey,
         assistant_id: assistantId,
+        filter,
+        source,
+        sort,
       },
     });
   }
+
+  sync(id) {
+    return axios.post(`${this.url}/${id}/sync`);
+  }
 }
 
 export default new CaptainDocument();
diff --git a/app/javascript/dashboard/components-next/captain/assistant/DocumentBulkActions.vue b/app/javascript/dashboard/components-next/captain/assistant/DocumentBulkActions.vue
new file mode 100644
index 000000000..378860b3e
--- /dev/null
+++ b/app/javascript/dashboard/components-next/captain/assistant/DocumentBulkActions.vue
@@ -0,0 +1,113 @@
+<script setup>
+import { computed, ref } from 'vue';
+import { useI18n } from 'vue-i18n';
+import { useStore } from 'dashboard/composables/store';
+import { useAlert } from 'dashboard/composables';
+
+import BulkSelectBar from 'dashboard/components-next/captain/assistant/BulkSelectBar.vue';
+import BulkDeleteDialog from 'dashboard/components-next/captain/pageComponents/BulkDeleteDialog.vue';
+import Button from 'dashboard/components-next/button/Button.vue';
+
+const props = defineProps({
+  selectedIds: { type: Set, default: () => new Set() },
+  documents: { type: Array, default: () => [] },
+});
+
+const emit = defineEmits([
+  'update:selectedIds',
+  'bulkSyncQueued',
+  'bulkDeleteSucceeded',
+]);
+
+const { t } = useI18n();
+const store = useStore();
+
+const bulkDeleteDialog = ref(null);
+
+const isSyncableDocument = doc =>
+  !doc.pdf_document && doc.status === 'available' && !doc.sync_in_progress;
+
+const syncableSelectedIds = computed(() => {
+  if (!props.selectedIds.size) return [];
+  return props.documents
+    .filter(doc => props.selectedIds.has(doc.id) && isSyncableDocument(doc))
+    .map(doc => doc.id);
+});
+
+const hasSyncableSelection = computed(
+  () => syncableSelectedIds.value.length > 0
+);
+
+const selectAllLabel = computed(() => {
+  const count = props.documents.length;
+  const isAllSelected = props.selectedIds.size === count && count > 0;
+  return isAllSelected
+    ? t('CAPTAIN.DOCUMENTS.UNSELECT_ALL', { count })
+    : t('CAPTAIN.DOCUMENTS.SELECT_ALL', { count });
+});
+
+const selectedCountLabel = computed(() =>
+  t('CAPTAIN.DOCUMENTS.SELECTED', { count: props.selectedIds.size })
+);
+
+const handleBulkSync = async () => {
+  const ids = syncableSelectedIds.value;
+  if (!ids.length) return;
+
+  try {
+    const response = await store.dispatch('captainBulkActions/handleBulkSync', {
+      ids,
+    });
+    const queuedCount = response?.count ?? response?.ids?.length ?? 0;
+    let message = t('CAPTAIN.DOCUMENTS.BULK_SYNC.ZERO_MESSAGE');
+
+    if (queuedCount === 1) {
+      message = t('CAPTAIN.DOCUMENTS.BULK_SYNC.SUCCESS_MESSAGE_ONE');
+    } else if (queuedCount > 1) {
+      message = t('CAPTAIN.DOCUMENTS.BULK_SYNC.SUCCESS_MESSAGE', {
+        count: queuedCount,
+      });
+    }
+
+    useAlert(message);
+    emit('update:selectedIds', new Set());
+    if (queuedCount > 0) emit('bulkSyncQueued');
+  } catch (error) {
+    useAlert(t('CAPTAIN.DOCUMENTS.BULK_SYNC.ERROR_MESSAGE'));
+  }
+};
+</script>
+
+<template>
+  <div>
+    <BulkSelectBar
+      :model-value="selectedIds"
+      :all-items="documents"
+      :select-all-label="selectAllLabel"
+      :selected-count-label="selectedCountLabel"
+      :delete-label="$t('CAPTAIN.DOCUMENTS.BULK_DELETE_BUTTON')"
+      class="w-fit"
+      :class="{ 'mb-2': selectedIds.size > 0 }"
+      @update:model-value="emit('update:selectedIds', $event)"
+      @bulk-delete="bulkDeleteDialog.dialogRef.open()"
+    >
+      <template v-if="hasSyncableSelection" #secondaryActions>
+        <Button
+          :label="$t('CAPTAIN.DOCUMENTS.BULK_SYNC_BUTTON')"
+          sm
+          slate
+          ghost
+          icon="i-lucide-refresh-cw"
+          class="!px-1.5"
+          @click="handleBulkSync"
+        />
+      </template>
+    </BulkSelectBar>
+    <BulkDeleteDialog
+      ref="bulkDeleteDialog"
+      :bulk-ids="selectedIds"
+      type="AssistantDocument"
+      @delete-success="emit('bulkDeleteSucceeded')"
+    />
+  </div>
+</template>
diff --git a/app/javascript/dashboard/components-next/captain/assistant/DocumentCard.vue b/app/javascript/dashboard/components-next/captain/assistant/DocumentCard.vue
index 30ebfc448..c8a011b9a 100644
--- a/app/javascript/dashboard/components-next/captain/assistant/DocumentCard.vue
+++ b/app/javascript/dashboard/components-next/captain/assistant/DocumentCard.vue
@@ -5,14 +5,17 @@ import { useI18n } from 'vue-i18n';
 import { dynamicTime } from 'shared/helpers/timeHelper';
 import { usePolicy } from 'dashboard/composables/usePolicy';
 import {
-  isPdfDocument,
+  isSafeHttpLink,
   formatDocumentLink,
+  getDocumentDisplayPath,
 } from 'shared/helpers/documentHelper';
 
+import Icon from 'dashboard/components-next/icon/Icon.vue';
 import CardLayout from 'dashboard/components-next/CardLayout.vue';
 import DropdownMenu from 'dashboard/components-next/dropdown-menu/DropdownMenu.vue';
 import Button from 'dashboard/components-next/button/Button.vue';
 import Checkbox from 'dashboard/components-next/checkbox/Checkbox.vue';
+import DocumentSyncStatus from 'dashboard/components-next/captain/assistant/DocumentSyncStatus.vue';
 
 const props = defineProps({
   id: {
@@ -31,10 +34,38 @@ const props = defineProps({
     type: String,
     required: true,
   },
+  pdfDocument: {
+    type: Boolean,
+    default: false,
+  },
   createdAt: {
     type: Number,
     required: true,
   },
+  status: {
+    type: String,
+    default: null,
+  },
+  syncStatus: {
+    type: String,
+    default: null,
+  },
+  lastSyncedAt: {
+    type: Number,
+    default: null,
+  },
+  lastSyncErrorCode: {
+    type: String,
+    default: null,
+  },
+  syncInProgress: {
+    type: Boolean,
+    default: false,
+  },
+  syncStaleAfterHours: {
+    type: Number,
+    default: null,
+  },
   isSelected: {
     type: Boolean,
     default: false,
@@ -64,6 +95,20 @@ const modelValue = computed({
   set: () => emit('select', props.id),
 });
 
+const isPdf = computed(() => props.pdfDocument);
+const hasSafeLink = computed(() => isSafeHttpLink(props.externalLink));
+const canManage = computed(() => checkPermissions(['administrator']));
+const isAvailable = computed(() => props.status === 'available');
+const canSync = computed(
+  () => canManage.value && !isPdf.value && isAvailable.value
+);
+const isSyncing = computed(() => props.syncStatus === 'syncing');
+const isFailed = computed(() => props.syncStatus === 'failed');
+const isRetryableSync = computed(
+  () => isFailed.value || (isSyncing.value && !props.syncInProgress)
+);
+const showSyncStatus = computed(() => !isPdf.value);
+
 const menuItems = computed(() => {
   const allOptions = [
     {
@@ -74,7 +119,19 @@ const menuItems = computed(() => {
     },
   ];
 
-  if (checkPermissions(['administrator'])) {
+  if (canSync.value) {
+    allOptions.push({
+      label: isRetryableSync.value
+        ? t('CAPTAIN.DOCUMENTS.OPTIONS.RETRY_SYNC')
+        : t('CAPTAIN.DOCUMENTS.OPTIONS.SYNC_NOW'),
+      value: 'sync',
+      action: 'sync',
+      icon: 'i-lucide-refresh-cw',
+      disabled: props.syncInProgress,
+    });
+  }
+
+  if (canManage.value) {
     allOptions.push({
       label: t('CAPTAIN.DOCUMENTS.OPTIONS.DELETE_DOCUMENT'),
       value: 'delete',
@@ -86,17 +143,25 @@ const menuItems = computed(() => {
   return allOptions;
 });
 
-const createdAt = computed(() => dynamicTime(props.createdAt));
+const createdAtLabel = computed(() => dynamicTime(props.createdAt));
 
-const displayLink = computed(() => formatDocumentLink(props.externalLink));
+const displayLink = computed(() =>
+  isPdf.value
+    ? formatDocumentLink(props.externalLink)
+    : getDocumentDisplayPath(props.externalLink)
+);
 const linkIcon = computed(() =>
-  isPdfDocument(props.externalLink) ? 'i-ph-file-pdf' : 'i-ph-link-simple'
+  isPdf.value ? 'i-ph-file-pdf' : 'i-ph-link-simple'
 );
 
 const handleAction = ({ action, value }) => {
   toggleDropdown(false);
   emit('action', { action, value, id: props.id });
 };
+
+const handleRetry = () => {
+  emit('action', { action: 'sync', id: props.id });
+};
 </script>
 
 <template>
@@ -141,17 +206,41 @@ const handleAction = ({ action, value }) => {
       <span
         class="flex gap-1 items-center text-sm truncate shrink-0 text-n-slate-11"
       >
-        <i class="i-woot-captain" />
+        <Icon icon="i-woot-captain" />
         {{ assistant?.name || '' }}
       </span>
+      <a
+        v-if="!isPdf && hasSafeLink"
+        :href="externalLink"
+        :title="externalLink"
+        target="_blank"
+        rel="noopener noreferrer"
+        class="flex flex-1 gap-1 justify-start items-center text-sm truncate text-n-slate-11 hover:text-n-slate-12 hover:underline"
+        @click.stop
+      >
+        <Icon :icon="linkIcon" class="shrink-0" />
+        <span class="truncate">{{ displayLink }}</span>
+        <Icon icon="i-lucide-external-link size-3 shrink-0 opacity-70" />
+      </a>
       <span
+        v-else
         class="flex flex-1 gap-1 justify-start items-center text-sm truncate text-n-slate-11"
       >
-        <i :class="linkIcon" class="shrink-0" />
+        <Icon :icon="linkIcon" class="shrink-0" />
         <span class="truncate">{{ displayLink }}</span>
       </span>
-      <div class="text-sm shrink-0 text-n-slate-11 line-clamp-1">
-        {{ createdAt }}
+      <DocumentSyncStatus
+        v-if="showSyncStatus"
+        :status="syncStatus"
+        :last-synced-at="lastSyncedAt"
+        :error-code="lastSyncErrorCode"
+        :sync-in-progress="syncInProgress"
+        :stale-after-hours="syncStaleAfterHours"
+        :show-retry="canSync && isRetryableSync"
+        @retry="handleRetry"
+      />
+      <div v-else class="text-sm shrink-0 text-n-slate-11 line-clamp-1">
+        {{ createdAtLabel }}
       </div>
     </div>
   </CardLayout>
diff --git a/app/javascript/dashboard/components-next/captain/assistant/DocumentFilter.vue b/app/javascript/dashboard/components-next/captain/assistant/DocumentFilter.vue
new file mode 100644
index 000000000..63b06daa7
--- /dev/null
+++ b/app/javascript/dashboard/components-next/captain/assistant/DocumentFilter.vue
@@ -0,0 +1,76 @@
+<script setup>
+import { computed, ref } from 'vue';
+import { debounce } from '@chatwoot/utils';
+
+import DocumentFiltersBar from 'dashboard/components-next/captain/assistant/DocumentFiltersBar.vue';
+
+const emit = defineEmits(['change']);
+
+const source = ref('all');
+const status = ref(null);
+const sort = ref('recently_updated');
+const query = ref('');
+
+const hasActiveFilters = computed(
+  () =>
+    source.value !== 'all' ||
+    Boolean(status.value) ||
+    Boolean(query.value.trim())
+);
+
+const buildParams = (page = 1) => {
+  const params = { page };
+  if (source.value !== 'all') params.source = source.value;
+  if (status.value) params.filter = status.value;
+  if (sort.value) params.sort = sort.value;
+  if (query.value.trim()) params.searchKey = query.value.trim();
+  return params;
+};
+
+const emitChange = () => emit('change');
+const debouncedEmitChange = debounce(emitChange, 300);
+
+const handleSourceSelect = sourceKey => {
+  source.value = sourceKey;
+  if (sourceKey !== 'web') status.value = null;
+  emitChange();
+};
+
+const handleStatusSelect = statusKey => {
+  status.value = statusKey;
+  if (statusKey) source.value = 'web';
+  emitChange();
+};
+
+const handleSortSelect = sortKey => {
+  sort.value = sortKey;
+  emitChange();
+};
+
+const handleSearch = value => {
+  query.value = value;
+  debouncedEmitChange();
+};
+
+const reset = () => {
+  source.value = 'all';
+  status.value = null;
+  sort.value = 'recently_updated';
+  query.value = '';
+};
+
+defineExpose({ buildParams, reset, hasActiveFilters });
+</script>
+
+<template>
+  <DocumentFiltersBar
+    :active-source-filter="source"
+    :active-status-filter="status"
+    :active-sort="sort"
+    :search-query="query"
+    @select-source="handleSourceSelect"
+    @select-status="handleStatusSelect"
+    @select-sort="handleSortSelect"
+    @search="handleSearch"
+  />
+</template>
diff --git a/app/javascript/dashboard/components-next/captain/assistant/DocumentFiltersBar.vue b/app/javascript/dashboard/components-next/captain/assistant/DocumentFiltersBar.vue
new file mode 100644
index 000000000..30d44b42e
--- /dev/null
+++ b/app/javascript/dashboard/components-next/captain/assistant/DocumentFiltersBar.vue
@@ -0,0 +1,165 @@
+<script setup>
+import { computed, ref } from 'vue';
+import { useI18n } from 'vue-i18n';
+
+import Icon from 'dashboard/components-next/icon/Icon.vue';
+import Input from 'dashboard/components-next/input/Input.vue';
+import Button from 'dashboard/components-next/button/Button.vue';
+import DropdownMenu from 'dashboard/components-next/dropdown-menu/DropdownMenu.vue';
+
+const props = defineProps({
+  activeSourceFilter: { type: String, default: 'all' },
+  activeStatusFilter: { type: String, default: null },
+  activeSort: { type: String, default: 'recently_updated' },
+  searchQuery: { type: String, default: '' },
+});
+
+const emit = defineEmits([
+  'selectSource',
+  'selectStatus',
+  'selectSort',
+  'search',
+]);
+
+const { t } = useI18n();
+
+const openMenu = ref(null);
+
+const MENU_CONFIG = [
+  {
+    key: 'source',
+    activeKey: 'activeSourceFilter',
+    dropdownClass: 'min-w-48',
+    options: [
+      { labelKey: 'SOURCE.ALL', value: 'all', icon: 'i-lucide-files' },
+      { labelKey: 'SOURCE.WEB', value: 'web', icon: 'i-lucide-link' },
+      { labelKey: 'SOURCE.PDF', value: 'pdf', icon: 'i-lucide-file-text' },
+    ],
+  },
+  {
+    key: 'status',
+    activeKey: 'activeStatusFilter',
+    dropdownClass: 'min-w-52',
+    options: [
+      { labelKey: 'STATUS.ANY', value: null, icon: 'i-lucide-circle-dashed' },
+      {
+        labelKey: 'STATUS.UPDATED',
+        value: 'synced',
+        icon: 'i-lucide-check-circle',
+      },
+      {
+        labelKey: 'STATUS.NEEDS_UPDATE',
+        value: 'stale',
+        icon: 'i-lucide-clock',
+      },
+      {
+        labelKey: 'STATUS.UPDATING',
+        value: 'syncing',
+        icon: 'i-lucide-refresh-cw',
+      },
+      {
+        labelKey: 'STATUS.FAILED',
+        value: 'failed',
+        icon: 'i-lucide-circle-x',
+      },
+    ],
+  },
+  {
+    key: 'sort',
+    activeKey: 'activeSort',
+    dropdownClass: 'min-w-56',
+    options: [
+      {
+        labelKey: 'SORT.RECENTLY_UPDATED',
+        value: 'recently_updated',
+        icon: 'i-lucide-arrow-down-up',
+      },
+      {
+        labelKey: 'SORT.RECENTLY_CREATED',
+        value: 'recently_created',
+        icon: 'i-lucide-clock',
+      },
+    ],
+  },
+];
+
+const filterMenus = computed(() =>
+  MENU_CONFIG.filter(
+    menu => !(menu.key === 'status' && props.activeSourceFilter === 'pdf')
+  ).map(menu => {
+    const active = props[menu.activeKey];
+    const items = menu.options.map(opt => ({
+      label: t(`CAPTAIN.DOCUMENTS.FILTERS.${opt.labelKey}`),
+      value: opt.value,
+      icon: opt.icon,
+      action: menu.key,
+      isSelected: opt.value === active,
+    }));
+    return {
+      ...menu,
+      items,
+      selected: items.find(item => item.isSelected) || items[0],
+    };
+  })
+);
+
+const closeMenu = () => {
+  openMenu.value = null;
+};
+
+const toggleMenu = menu => {
+  openMenu.value = openMenu.value === menu ? null : menu;
+};
+
+const handleMenuAction = ({ action, value }) => {
+  closeMenu();
+  if (action === 'source') emit('selectSource', value);
+  else if (action === 'status') emit('selectStatus', value);
+  else if (action === 'sort') emit('selectSort', value);
+};
+</script>
+
+<template>
+  <div
+    v-on-clickaway="closeMenu"
+    class="flex flex-col gap-3 w-full lg:flex-row lg:items-center lg:justify-between"
+  >
+    <div class="flex flex-wrap items-center gap-2">
+      <div v-for="menu in filterMenus" :key="menu.key" class="relative">
+        <Button
+          :icon="menu.selected.icon"
+          slate
+          outline
+          :class="{ 'bg-n-slate-9/10': openMenu === menu.key }"
+          @click="toggleMenu(menu.key)"
+        >
+          <span class="min-w-0 truncate">{{ menu.selected.label }}</span>
+          <Icon icon="i-lucide-chevron-down" class="shrink-0 size-4" />
+        </Button>
+        <DropdownMenu
+          v-if="openMenu === menu.key"
+          :menu-items="menu.items"
+          :class="menu.dropdownClass"
+          class="top-full mt-2 ltr:left-0 rtl:right-0"
+          @action="handleMenuAction"
+        />
+      </div>
+    </div>
+
+    <Input
+      :model-value="searchQuery"
+      type="search"
+      :placeholder="t('CAPTAIN.DOCUMENTS.FILTERS.SEARCH_PLACEHOLDER')"
+      :custom-input-class="['ltr:!pl-9 rtl:!pr-9']"
+      class="w-full lg:max-w-72"
+      @input="emit('search', $event.target.value)"
+    >
+      <template #prefix>
+        <Icon
+          icon="i-lucide-search"
+          class="absolute size-4 text-n-slate-11 top-1/2 -translate-y-1/2 ltr:left-3 rtl:right-3"
+        />
+      </template>
+    </Input>
+  </div>
+</template>
diff --git a/app/javascript/dashboard/components-next/captain/assistant/DocumentSyncStatus.vue b/app/javascript/dashboard/components-next/captain/assistant/DocumentSyncStatus.vue
new file mode 100644
index 000000000..df0c99851
--- /dev/null
+++ b/app/javascript/dashboard/components-next/captain/assistant/DocumentSyncStatus.vue
@@ -0,0 +1,153 @@
+<script setup>
+import { computed } from 'vue';
+import { useI18n } from 'vue-i18n';
+import { dynamicTime } from 'shared/helpers/timeHelper';
+import Icon from 'dashboard/components-next/icon/Icon.vue';
+import Button from 'dashboard/components-next/button/Button.vue';
+import Spinner from 'dashboard/components-next/spinner/Spinner.vue';
+
+const props = defineProps({
+  status: {
+    type: String,
+    default: null,
+  },
+  lastSyncedAt: {
+    type: Number,
+    default: null,
+  },
+  errorCode: {
+    type: String,
+    default: null,
+  },
+  syncInProgress: {
+    type: Boolean,
+    default: false,
+  },
+  staleAfterHours: {
+    type: Number,
+    default: null,
+  },
+  showRetry: {
+    type: Boolean,
+    default: false,
+  },
+});
+
+const emit = defineEmits(['retry']);
+const { t } = useI18n();
+
+const SECONDS_PER_HOUR = 3600;
+
+const SYNCING = 'syncing';
+const FAILED = 'failed';
+
+const ERROR_CODE_LABELS = {
+  not_found: 'CAPTAIN.DOCUMENTS.SYNC_ERRORS.NOT_FOUND',
+  access_denied: 'CAPTAIN.DOCUMENTS.SYNC_ERRORS.ACCESS_DENIED',
+  timeout: 'CAPTAIN.DOCUMENTS.SYNC_ERRORS.TIMEOUT',
+  content_empty: 'CAPTAIN.DOCUMENTS.SYNC_ERRORS.CONTENT_EMPTY',
+  fetch_failed: 'CAPTAIN.DOCUMENTS.SYNC_ERRORS.FETCH_FAILED',
+  sync_error: 'CAPTAIN.DOCUMENTS.SYNC_ERRORS.SYNC_ERROR',
+};
+const DEFAULT_ERROR_LABEL = 'CAPTAIN.DOCUMENTS.SYNC_ERRORS.DEFAULT';
+
+const hasSyncingStatus = computed(() => props.status === SYNCING);
+const isSyncing = computed(
+  () => hasSyncingStatus.value && props.syncInProgress
+);
+const isStaleSync = computed(
+  () => hasSyncingStatus.value && !props.syncInProgress
+);
+const isFailed = computed(() => props.status === FAILED);
+const canRetry = computed(() => isFailed.value || isStaleSync.value);
+const hasBeenSynced = computed(() => Boolean(props.lastSyncedAt));
+
+const ageInHours = computed(() => {
+  if (!props.lastSyncedAt) return null;
+  const nowSeconds = Date.now() / 1000;
+  return (nowSeconds - props.lastSyncedAt) / SECONDS_PER_HOUR;
+});
+
+const staleAfterHours = computed(() => Number(props.staleAfterHours));
+const hasStaleThreshold = computed(
+  () => Number.isFinite(staleAfterHours.value) && staleAfterHours.value > 0
+);
+const isStale = computed(
+  () =>
+    hasStaleThreshold.value &&
+    ageInHours.value !== null &&
+    ageInHours.value >= staleAfterHours.value
+);
+
+const errorLabel = computed(() =>
+  t(ERROR_CODE_LABELS[props.errorCode] || DEFAULT_ERROR_LABEL)
+);
+
+const label = computed(() => {
+  if (isSyncing.value) return t('CAPTAIN.DOCUMENTS.SYNC_STATUS.SYNCING');
+  if (isStaleSync.value) return t('CAPTAIN.DOCUMENTS.SYNC_STATUS.STALE_SYNC');
+  if (isFailed.value) return t('CAPTAIN.DOCUMENTS.SYNC_STATUS.FAILED');
+  if (hasBeenSynced.value)
+    return t('CAPTAIN.DOCUMENTS.SYNC_STATUS.SYNCED', {
+      time: dynamicTime(props.lastSyncedAt),
+    });
+  return t('CAPTAIN.DOCUMENTS.SYNC_STATUS.NEVER_SYNCED');
+});
+
+const fullLabel = computed(() => {
+  if (isSyncing.value) return t('CAPTAIN.DOCUMENTS.SYNC_STATUS.SYNCING');
+  if (isStaleSync.value) return t('CAPTAIN.DOCUMENTS.SYNC_STATUS.STALE_SYNC');
+  if (isFailed.value)
+    return t('CAPTAIN.DOCUMENTS.SYNC_STATUS.FAILED', {
+      error: errorLabel.value,
+    });
+  if (hasBeenSynced.value)
+    return t('CAPTAIN.DOCUMENTS.SYNC_STATUS.SYNCED', {
+      time: dynamicTime(props.lastSyncedAt),
+    });
+  return t('CAPTAIN.DOCUMENTS.SYNC_STATUS.NEVER_SYNCED');
+});
+
+const tone = computed(() => {
+  if (isSyncing.value) return 'amber';
+  if (isStaleSync.value) return 'amber';
+  if (isFailed.value) return 'ruby';
+  if (isStale.value) return 'amber';
+  return 'slate';
+});
+
+const textClass = computed(() => {
+  if (tone.value === 'amber') return 'text-n-amber-11';
+  if (tone.value === 'ruby') return 'text-n-ruby-11';
+  return 'text-n-slate-11';
+});
+
+const statusIcon = computed(() => {
+  if (isFailed.value || isStale.value || isStaleSync.value) {
+    return 'i-lucide-circle-alert';
+  }
+  return 'i-lucide-refresh-cw';
+});
+</script>
+
+<template>
+  <span
+    class="flex gap-1.5 items-center text-sm truncate shrink-0 tabular-nums"
+    :class="textClass"
+    :title="fullLabel"
+  >
+    <Spinner v-if="isSyncing" class="text-n-amber-11 size-3" />
+    <Icon v-else :icon="statusIcon" class="shrink-0 size-3.5" />
+    <span class="truncate">{{ label }}</span>
+    <Button
+      v-if="showRetry && canRetry"
+      :label="t('CAPTAIN.DOCUMENTS.OPTIONS.RETRY_SYNC')"
+      xs
+      link
+      ruby
+      icon="i-lucide-refresh-cw"
+      class="hover:!no-underline !gap-1 ms-1"
+      @click.stop="emit('retry')"
+    />
+  </span>
+</template>
diff --git a/app/javascript/dashboard/components-next/captain/pageComponents/document/CreateDocumentDialog.vue b/app/javascript/dashboard/components-next/captain/pageComponents/document/CreateDocumentDialog.vue
index 05fc69342..ffdc95e25 100644
--- a/app/javascript/dashboard/components-next/captain/pageComponents/document/CreateDocumentDialog.vue
+++ b/app/javascript/dashboard/components-next/captain/pageComponents/document/CreateDocumentDialog.vue
@@ -15,7 +15,7 @@ defineProps({
   },
 });
 
-const emit = defineEmits(['close']);
+const emit = defineEmits(['close', 'createSuccess']);
 const { t } = useI18n();
 const store = useStore();
 
@@ -26,6 +26,7 @@ const i18nKey = 'CAPTAIN.DOCUMENTS.CREATE';
 const handleSubmit = async newDocument => {
   try {
     await store.dispatch('captainDocuments/create', newDocument);
+    emit('createSuccess');
     useAlert(t(`${i18nKey}.SUCCESS_MESSAGE`));
     dialogRef.value.close();
   } catch (error) {
diff --git a/app/javascript/dashboard/featureFlags.js b/app/javascript/dashboard/featureFlags.js
index 1f9632425..da296bac4 100644
--- a/app/javascript/dashboard/featureFlags.js
+++ b/app/javascript/dashboard/featureFlags.js
@@ -41,6 +41,7 @@ export const FEATURE_FLAGS = {
   CAPTAIN_CUSTOM_TOOLS: 'custom_tools',
   CAPTAIN_V2: 'captain_integration_v2',
   CAPTAIN_TASKS: 'captain_tasks',
+  CAPTAIN_DOCUMENT_AUTO_SYNC: 'captain_document_auto_sync',
   SAML: 'saml',
   QUOTED_EMAIL_REPLY: 'quoted_email_reply',
   COMPANIES: 'companies',
diff --git a/app/javascript/dashboard/i18n/locale/en/integrations.json b/app/javascript/dashboard/i18n/locale/en/integrations.json
index ad89755e1..6bf332b25 100644
--- a/app/javascript/dashboard/i18n/locale/en/integrations.json
+++ b/app/javascript/dashboard/i18n/locale/en/integrations.json
@@ -742,6 +742,7 @@
       "SELECT_ALL": "Select all ({count})",
       "UNSELECT_ALL": "Unselect all ({count})",
       "BULK_DELETE_BUTTON": "Delete",
+      "BULK_SYNC_BUTTON": "Refresh",
       "BULK_DELETE": {
         "TITLE": "Delete documents?",
         "DESCRIPTION": "Are you sure you want to delete the selected documents? This action cannot be undone.",
@@ -749,6 +750,51 @@
         "SUCCESS_MESSAGE": "Documents deleted successfully",
         "ERROR_MESSAGE": "There was an error deleting the documents, please try again."
       },
+      "BULK_SYNC": {
+        "SUCCESS_MESSAGE_ONE": "Refresh queued for 1 document",
+        "SUCCESS_MESSAGE": "Refresh queued for {count} documents",
+        "ZERO_MESSAGE": "No documents marked for refresh.",
+        "ERROR_MESSAGE": "There was an error queuing the refresh, please try again."
+      },
+      "SYNC": {
+        "QUEUED_MESSAGE": "Refresh queued. We'll update the document shortly.",
+        "ERROR_MESSAGE": "Could not queue refresh, please try again."
+      },
+      "FILTERS": {
+        "SOURCE": {
+          "ALL": "All sources",
+          "WEB": "Web pages",
+          "PDF": "PDFs"
+        },
+        "STATUS": {
+          "ANY": "Any status",
+          "UPDATED": "Updated",
+          "NEEDS_UPDATE": "Needs update",
+          "UPDATING": "Updating",
+          "FAILED": "Failed"
+        },
+        "SORT": {
+          "RECENTLY_UPDATED": "Recently updated",
+          "RECENTLY_CREATED": "Recently created"
+        },
+        "SEARCH_PLACEHOLDER": "Search..."
+      },
+      "SYNC_STATUS": {
+        "SYNCED": "last updated {time}",
+        "SYNCING": "updating...",
+        "STALE_SYNC": "update stalled",
+        "FAILED": "Failed to sync",
+        "NEVER_SYNCED": "not updated yet"
+      },
+      "SYNC_ERRORS": {
+        "NOT_FOUND": "Page not found",
+        "ACCESS_DENIED": "Access denied",
+        "TIMEOUT": "Page took too long to respond",
+        "CONTENT_EMPTY": "Page returned empty content",
+        "FETCH_FAILED": "Could not fetch page",
+        "SYNC_ERROR": "Unexpected error",
+        "DEFAULT": "Sync error"
+      },
       "RELATED_RESPONSES": {
         "TITLE": "Related FAQs",
         "DESCRIPTION": "These FAQs are generated directly from the document."
@@ -793,11 +839,15 @@
 
       "OPTIONS": {
         "VIEW_RELATED_RESPONSES": "View Related Responses",
+        "SYNC_NOW": "Refresh now",
+        "RETRY_SYNC": "Retry refresh",
         "DELETE_DOCUMENT": "Delete Document"
       },
       "EMPTY_STATE": {
         "TITLE": "No documents available",
         "SUBTITLE": "Documents are used by your assistant to generate FAQs. You can import documents to provide context for your assistant.",
+        "FILTERED_TITLE": "No matching documents",
+        "FILTERED_SUBTITLE": "Try changing the source, status, or search term.",
         "FEATURE_SPOTLIGHT": {
           "TITLE": "Captain Document",
           "NOTE": "A document in Captain serves as a knowledge resource for the assistant. By connecting your help center or guides, Captain can analyze the content and provide accurate responses for customer inquiries."
diff --git a/app/javascript/dashboard/routes/dashboard/captain/documents/Index.vue b/app/javascript/dashboard/routes/dashboard/captain/documents/Index.vue
index 2bd7446cb..e13207416 100644
--- a/app/javascript/dashboard/routes/dashboard/captain/documents/Index.vue
+++ b/app/javascript/dashboard/routes/dashboard/captain/documents/Index.vue
@@ -1,15 +1,17 @@
 <script setup>
-import { computed, onMounted, ref, nextTick } from 'vue';
+import { computed, onUnmounted, ref, nextTick, watch } from 'vue';
+import { useTimeoutPoll } from '@vueuse/core';
 import { useMapGetter, useStore } from 'dashboard/composables/store';
 import { useRoute } from 'vue-router';
 import { FEATURE_FLAGS } from 'dashboard/featureFlags';
 import { useAccount } from 'dashboard/composables/useAccount';
+import { useAlert } from 'dashboard/composables';
 import { usePolicy } from 'dashboard/composables/usePolicy';
 
 import DeleteDialog from 'dashboard/components-next/captain/pageComponents/DeleteDialog.vue';
 import DocumentCard from 'dashboard/components-next/captain/assistant/DocumentCard.vue';
-import BulkSelectBar from 'dashboard/components-next/captain/assistant/BulkSelectBar.vue';
-import BulkDeleteDialog from 'dashboard/components-next/captain/pageComponents/BulkDeleteDialog.vue';
+import DocumentFilter from 'dashboard/components-next/captain/assistant/DocumentFilter.vue';
+import DocumentBulkActions from 'dashboard/components-next/captain/assistant/DocumentBulkActions.vue';
 import Policy from 'dashboard/components/policy.vue';
 import PageLayout from 'dashboard/components-next/captain/PageLayout.vue';
 import CaptainPaywall from 'dashboard/components-next/captain/pageComponents/Paywall.vue';
@@ -18,6 +20,7 @@ import CreateDocumentDialog from 'dashboard/components-next/captain/pageComponen
 import DocumentPageEmptyState from 'dashboard/components-next/captain/pageComponents/emptyStates/DocumentPageEmptyState.vue';
 import FeatureSpotlightPopover from 'dashboard/components-next/feature-spotlight/FeatureSpotlightPopover.vue';
 import LimitBanner from 'dashboard/components-next/captain/pageComponents/document/LimitBanner.vue';
+import CaptainDocumentAPI from 'dashboard/api/captain/document';
 import { useI18n } from 'vue-i18n';
 
 const route = useRoute();
@@ -25,6 +28,9 @@ const store = useStore();
 const { t } = useI18n();
 const { checkPermissions } = usePolicy();
 
+const SYNC_POLL_INTERVAL_MS = 5000;
+const SYNC_POLL_MAX_DURATION_MS = 15 * 60 * 1000;
+
 const { isOnChatwootCloud } = useAccount();
 const uiFlags = useMapGetter('captainDocuments/getUIFlags');
 const documents = useMapGetter('captainDocuments/getRecords');
@@ -36,7 +42,6 @@ const canManageDocuments = computed(() => checkPermissions(['administrator']));
 
 const selectedDocument = ref(null);
 const deleteDocumentDialog = ref(null);
-const bulkDeleteDialog = ref(null);
 const bulkSelectedIds = ref(new Set());
 const hoveredCard = ref(null);
 
@@ -66,6 +71,152 @@ const handleCreateDialogClose = () => {
   showCreateDialog.value = false;
 };
 
+const documentFilter = ref(null);
+const syncIntervalHours = ref(null);
+
+const currentAssistantId = () =>
+  Number.isFinite(selectedAssistantId.value) ? selectedAssistantId.value : null;
+
+const buildDocumentFilterParams = (page = 1) => {
+  const filterParams = documentFilter.value?.buildParams(page) ?? {
+    page,
+    sort: 'recently_updated',
+  };
+  const assistantId = currentAssistantId();
+  if (assistantId) filterParams.assistantId = assistantId;
+  return filterParams;
+};
+
+let documentsRequestId = 0;
+let fetchingListRequestId = null;
+
+const isCurrentDocumentRequest = (requestId, filterParams) =>
+  requestId === documentsRequestId &&
+  (filterParams.assistantId || null) === currentAssistantId();
+
+const pruneSelectionToDocuments = nextDocuments => {
+  if (!bulkSelectedIds.value.size) return;
+
+  const visibleDocumentIds = new Set(nextDocuments.map(doc => doc.id));
+  const selectedIds = new Set(
+    [...bulkSelectedIds.value].filter(id => visibleDocumentIds.has(id))
+  );
+
+  if (selectedIds.size !== bulkSelectedIds.value.size) {
+    bulkSelectedIds.value = selectedIds;
+  }
+};
+
+const fetchDocuments = async (page = 1, { showLoader = true } = {}) => {
+  documentsRequestId += 1;
+  const requestId = documentsRequestId;
+  const filterParams = buildDocumentFilterParams(page);
+
+  if (showLoader) {
+    fetchingListRequestId = requestId;
+    store.dispatch('captainDocuments/setFetchingList', true);
+  }
+
+  try {
+    const response = await CaptainDocumentAPI.get(filterParams);
+
+    if (!isCurrentDocumentRequest(requestId, filterParams)) {
+      return [];
+    }
+
+    const { payload, meta } = response.data;
+    store.dispatch('captainDocuments/setRecords', { records: payload, meta });
+    pruneSelectionToDocuments(payload);
+    syncIntervalHours.value = Number(meta?.sync_interval_hours) || null;
+    return payload;
+  } catch (error) {
+    if (isCurrentDocumentRequest(requestId, filterParams)) {
+      throw error;
+    }
+    return [];
+  } finally {
+    if (showLoader && fetchingListRequestId === requestId) {
+      fetchingListRequestId = null;
+      store.dispatch('captainDocuments/setFetchingList', false);
+    }
+  }
+};
+
+const refreshDocumentsPage = (
+  page = documentsMeta.value?.page || 1,
+  { showLoader = false } = {}
+) => {
+  return fetchDocuments(page, { showLoader }).catch(() => {});
+};
+
+const onFiltersChanged = () => {
+  bulkSelectedIds.value = new Set();
+  fetchDocuments(1);
+};
+
+const syncPollStartedAt = ref(null);
+
+const hasDocumentsSyncing = computed(() =>
+  (documents.value || []).some(doc => doc.sync_in_progress)
+);
+
+const hasSyncingDocuments = computed(() => hasDocumentsSyncing.value);
+
+const isWithinSyncPollWindow = () =>
+  syncPollStartedAt.value &&
+  Date.now() - syncPollStartedAt.value < SYNC_POLL_MAX_DURATION_MS;
+
+const shouldContinueSyncPolling = () =>
+  hasSyncingDocuments.value && isWithinSyncPollWindow();
+
+let syncPollingControls;
+
+function stopSyncPolling() {
+  syncPollingControls.pause();
+  syncPollStartedAt.value = null;
+}
+
+async function pollSyncDocuments() {
+  try {
+    await refreshDocumentsPage();
+  } catch (error) {
+    // Keep the existing polling decision based on the last known sync state.
+  }
+
+  if (!shouldContinueSyncPolling()) {
+    stopSyncPolling();
+  }
+}
+
+function scheduleSyncPoll({ extendWindow = false } = {}) {
+  if (extendWindow || !syncPollStartedAt.value) {
+    syncPollStartedAt.value = Date.now();
+  }
+
+  if (syncPollingControls.isActive.value) return;
+  syncPollingControls.resume();
+}
+
+syncPollingControls = useTimeoutPoll(pollSyncDocuments, SYNC_POLL_INTERVAL_MS, {
+  immediate: false,
+});
+
+watch(hasSyncingDocuments, isSyncing => {
+  if (isSyncing) {
+    scheduleSyncPoll();
+  }
+});
+
+const handleSync = async id => {
+  try {
+    await store.dispatch('captainDocuments/sync', id);
+    useAlert(t('CAPTAIN.DOCUMENTS.SYNC.QUEUED_MESSAGE'));
+    scheduleSyncPoll({ extendWindow: true });
+  } catch (error) {
+    useAlert(t('CAPTAIN.DOCUMENTS.SYNC.ERROR_MESSAGE'));
+  }
+};
+
 const handleAction = ({ action, id }) => {
   selectedDocument.value = documents.value.find(
     captainDocument => id === captainDocument.id
@@ -76,19 +227,12 @@ const handleAction = ({ action, id }) => {
       handleDelete();
     } else if (action === 'viewRelatedQuestions') {
       handleShowRelatedDocument();
+    } else if (action === 'sync') {
+      handleSync(id);
     }
   });
 };
 
-const fetchDocuments = (page = 1) => {
-  const filterParams = { page };
-
-  if (selectedAssistantId.value) {
-    filterParams.assistantId = selectedAssistantId.value;
-  }
-  store.dispatch('captainDocuments/get', filterParams);
-};
-
 const onPageChange = page => {
   const hadSelection = bulkSelectedIds.value.size > 0;
   fetchDocuments(page);
@@ -101,31 +245,14 @@ const onPageChange = page => {
 const onDeleteSuccess = () => {
   if (documents.value?.length === 0 && documentsMeta.value?.page > 1) {
     onPageChange(documentsMeta.value.page - 1);
+  } else {
+    refreshDocumentsPage();
   }
 };
 
-const buildSelectedCountLabel = computed(() => {
-  const count = documents.value?.length || 0;
-  const isAllSelected = bulkSelectedIds.value.size === count && count > 0;
-  return isAllSelected
-    ? t('CAPTAIN.DOCUMENTS.UNSELECT_ALL', { count })
-    : t('CAPTAIN.DOCUMENTS.SELECT_ALL', { count });
-});
-
-const selectedCountLabel = computed(() => {
-  return t('CAPTAIN.DOCUMENTS.SELECTED', {
-    count: bulkSelectedIds.value.size,
-  });
-});
-
-const hasBulkSelection = computed(() => bulkSelectedIds.value.size > 0);
-
-const shouldShowSelectionControl = docId => {
-  return (
-    canManageDocuments.value &&
-    (hoveredCard.value === docId || hasBulkSelection.value)
-  );
-};
+const shouldShowSelectionControl = docId =>
+  canManageDocuments.value &&
+  (hoveredCard.value === docId || bulkSelectedIds.value.size > 0);
 
 const handleCardHover = (isHovered, id) => {
   hoveredCard.value = isHovered ? id : null;
@@ -152,12 +279,30 @@ const fetchDocumentsAfterBulkAction = () => {
   bulkSelectedIds.value = new Set();
 };
 
-const onBulkDeleteSuccess = () => {
-  fetchDocumentsAfterBulkAction();
+const onCreateSuccess = () => {
+  refreshDocumentsPage(1);
 };
 
-onMounted(() => {
-  fetchDocuments();
+const hasActiveDocumentFilters = computed(
+  () => documentFilter.value?.hasActiveFilters ?? false
+);
+
+watch(
+  selectedAssistantId,
+  async () => {
+    documentFilter.value?.reset();
+    bulkSelectedIds.value = new Set();
+    syncIntervalHours.value = null;
+    stopSyncPolling();
+    await fetchDocuments(1);
+    if (hasSyncingDocuments.value) scheduleSyncPoll();
+  },
+  { immediate: true }
+);
+
+onUnmounted(() => {
+  stopSyncPolling();
+  documentsRequestId += 1;
 });
 </script>
 
@@ -170,27 +315,27 @@ onMounted(() => {
     :current-page="documentsMeta.page"
     :show-pagination-footer="!isFetching && !!documents.length"
     :is-fetching="isFetching"
-    :is-empty="!documents.length"
-    :show-know-more="false"
+    :is-empty="!documents.length && !hasActiveDocumentFilters"
     :feature-flag="FEATURE_FLAGS.CAPTAIN"
     @update:current-page="onPageChange"
     @click="handleCreateDocument"
   >
     <template #subHeader>
       <Policy :permissions="['administrator']">
-        <BulkSelectBar
-          v-model="bulkSelectedIds"
-          :all-items="documents"
-          :select-all-label="buildSelectedCountLabel"
-          :selected-count-label="selectedCountLabel"
-          :delete-label="$t('CAPTAIN.DOCUMENTS.BULK_DELETE_BUTTON')"
-          class="w-fit"
-          :class="{ 'mb-2': bulkSelectedIds.size > 0 }"
-          @bulk-delete="bulkDeleteDialog.dialogRef.open()"
+        <DocumentBulkActions
+          v-model:selected-ids="bulkSelectedIds"
+          :documents="documents"
+          @bulk-sync-queued="scheduleSyncPoll({ extendWindow: true })"
+          @bulk-delete-succeeded="fetchDocumentsAfterBulkAction"
         />
       </Policy>
+      <DocumentFilter
+        v-show="!bulkSelectedIds.size"
+        ref="documentFilter"
+        class="mb-2"
+        @change="onFiltersChanged"
+      />
     </template>
-
     <template #knowMore>
       <FeatureSpotlightPopover
         :button-label="$t('CAPTAIN.HEADER_KNOW_MORE')"
@@ -214,15 +359,34 @@ onMounted(() => {
     <template #body>
       <LimitBanner class="mb-5" />
 
-      <div class="flex flex-col gap-4">
+      <div
+        v-if="!documents.length && hasActiveDocumentFilters"
+        class="flex flex-col items-center justify-center min-h-80 gap-2 text-center"
+      >
+        <span class="text-base font-medium text-n-slate-12">
+          {{ $t('CAPTAIN.DOCUMENTS.EMPTY_STATE.FILTERED_TITLE') }}
+        </span>
+        <span class="max-w-md text-sm text-n-slate-11">
+          {{ $t('CAPTAIN.DOCUMENTS.EMPTY_STATE.FILTERED_SUBTITLE') }}
+        </span>
+      </div>
+
+      <div v-else class="flex flex-col gap-4">
         <DocumentCard
           v-for="doc in documents"
           :id="doc.id"
           :key="doc.id"
           :name="doc.name || doc.external_link"
           :external-link="doc.external_link"
+          :pdf-document="doc.pdf_document"
           :assistant="doc.assistant"
           :created-at="doc.created_at"
+          :status="doc.status"
+          :sync-status="doc.sync_status"
+          :last-synced-at="doc.last_synced_at"
+          :last-sync-error-code="doc.last_sync_error_code"
+          :sync-in-progress="doc.sync_in_progress"
+          :sync-stale-after-hours="syncIntervalHours"
           :is-selected="canManageDocuments && bulkSelectedIds.has(doc.id)"
           :selectable="canManageDocuments"
           :show-selection-control="shouldShowSelectionControl(doc.id)"
@@ -244,6 +408,7 @@ onMounted(() => {
       v-if="showCreateDialog"
       ref="createDocumentDialog"
       :assistant-id="selectedAssistantId"
+      @create-success="onCreateSuccess"
       @close="handleCreateDialogClose"
     />
     <DeleteDialog
@@ -253,12 +418,5 @@ onMounted(() => {
       type="Documents"
       @delete-success="onDeleteSuccess"
     />
-    <BulkDeleteDialog
-      v-if="bulkSelectedIds"
-      ref="bulkDeleteDialog"
-      :bulk-ids="bulkSelectedIds"
-      type="AssistantDocument"
-      @delete-success="onBulkDeleteSuccess"
-    />
   </PageLayout>
 </template>
diff --git a/app/javascript/dashboard/store/captain/bulkActions.js b/app/javascript/dashboard/store/captain/bulkActions.js
index 436801092..1fab006f6 100644
--- a/app/javascript/dashboard/store/captain/bulkActions.js
+++ b/app/javascript/dashboard/store/captain/bulkActions.js
@@ -61,5 +61,18 @@ export default createStore({
       });
       return response;
     },
+
+    handleBulkSync: async function handleBulkSync({ dispatch }, { ids }) {
+      const response = await dispatch('processBulkAction', {
+        type: 'AssistantDocument',
+        actionType: 'sync',
+        ids,
+      });
+
+      await dispatch('captainDocuments/markSyncing', response.ids || [], {
+        root: true,
+      });
+      return response;
+    },
   }),
 });
diff --git a/app/javascript/dashboard/store/captain/document.js b/app/javascript/dashboard/store/captain/document.js
index 76d0c9124..fbe017e26 100644
--- a/app/javascript/dashboard/store/captain/document.js
+++ b/app/javascript/dashboard/store/captain/document.js
@@ -1,15 +1,55 @@
 import CaptainDocumentAPI from 'dashboard/api/captain/document';
+import { throwErrorMessage } from 'dashboard/store/utils/api';
 import { createStore } from '../storeFactory';
 
+const SYNCING_STATE = 'syncing';
+
+const markRecordsSyncing = (records, ids) => {
+  const idSet = new Set(ids);
+  return records.map(record =>
+    idSet.has(record.id)
+      ? {
+          ...record,
+          sync_status: SYNCING_STATE,
+          sync_in_progress: true,
+          last_sync_attempted_at: Math.floor(Date.now() / 1000),
+          last_sync_error_code: null,
+        }
+      : record
+  );
+};
+
 export default createStore({
   name: 'CaptainDocument',
   API: CaptainDocumentAPI,
+  getters: {
+    getRecords: state => state.records,
+  },
   actions: mutations => ({
+    setFetchingList({ commit }, isFetching) {
+      commit(mutations.SET_UI_FLAG, { fetchingList: isFetching });
+    },
+    setRecords({ commit }, { records, meta }) {
+      commit(mutations.SET, records);
+      commit(mutations.SET_META, meta);
+    },
     removeBulkRecords({ commit, getters }, ids) {
       const records = getters.getRecords.filter(
         record => !ids.includes(record.id)
       );
       commit(mutations.SET, records);
     },
+    markSyncing({ commit, getters }, ids) {
+      commit(mutations.SET, markRecordsSyncing(getters.getRecords, ids));
+    },
+    async sync({ dispatch }, id) {
+      try {
+        await CaptainDocumentAPI.sync(id);
+        dispatch('markSyncing', [id]);
+        return id;
+      } catch (error) {
+        return throwErrorMessage(error);
+      }
+    },
   }),
 });
diff --git a/app/javascript/shared/helpers/documentHelper.js b/app/javascript/shared/helpers/documentHelper.js
index f050ad54d..ca7610f53 100644
--- a/app/javascript/shared/helpers/documentHelper.js
+++ b/app/javascript/shared/helpers/documentHelper.js
@@ -5,6 +5,7 @@
 // Constants for document processing
 const PDF_PREFIX = 'PDF:';
 const TIMESTAMP_PATTERN = /_\d{14}(?=\.pdf$)/; // Format: _YYYYMMDDHHMMSS before .pdf extension
+const URL_DISPLAY_PREFIX_PATTERN = /^https?:\/\/(www\.)?/i;
 
 /**
  * Checks if a document is a PDF based on its external link
@@ -16,10 +17,26 @@ export const isPdfDocument = externalLink => {
   return externalLink.startsWith(PDF_PREFIX);
 };
 
+/**
+ * Checks if a link is safe to bind to an href attribute (http/https only).
+ * Guards against schemes like `javascript:` that would execute on click.
+ * @param {string} externalLink - The external link string
+ * @returns {boolean} True if the link uses http or https
+ */
+export const isSafeHttpLink = externalLink => {
+  if (!externalLink) return false;
+  try {
+    const { protocol } = new URL(externalLink);
+    return protocol === 'http:' || protocol === 'https:';
+  } catch (e) {
+    return false;
+  }
+};
+
 /**
  * Formats the display link for documents
  * For PDF documents: removes 'PDF:' prefix and timestamp suffix
- * For regular URLs: returns as-is
+ * For regular URLs: strips http(s):// and www. for a denser list view
  *
  * @param {string} externalLink - The external link string
  * @returns {string} Formatted display link
@@ -34,5 +51,28 @@ export const formatDocumentLink = externalLink => {
     return fullName.replace(TIMESTAMP_PATTERN, '');
   }
 
-  return externalLink;
+  return externalLink.replace(URL_DISPLAY_PREFIX_PATTERN, '');
+};
+
+/**
+ * Returns the path of a URL for compact display in document lists. This avoids
+ * repeating the domain while preserving enough context to distinguish pages.
+ * Falls back to the bare hostname for root URLs and formatDocumentLink for
+ * malformed URLs and PDFs.
+ */
+export const getDocumentDisplayPath = externalLink => {
+  if (!externalLink) return '';
+  if (isPdfDocument(externalLink)) return formatDocumentLink(externalLink);
+  try {
+    const { pathname, hostname } = new URL(externalLink);
+    const path = pathname.replace(/^\/+/, '');
+    if (!path) return hostname.replace(/^www\./i, '');
+    try {
+      return decodeURIComponent(path);
+    } catch (e) {
+      return path;
+    }
+  } catch (e) {
+    return formatDocumentLink(externalLink);
+  }
 };
diff --git a/app/javascript/shared/helpers/specs/documentHelper.spec.js b/app/javascript/shared/helpers/specs/documentHelper.spec.js
index 64baf7069..dc01e078d 100644
--- a/app/javascript/shared/helpers/specs/documentHelper.spec.js
+++ b/app/javascript/shared/helpers/specs/documentHelper.spec.js
@@ -1,5 +1,6 @@
 import {
   isPdfDocument,
+  isSafeHttpLink,
   formatDocumentLink,
 } from 'shared/helpers/documentHelper';
 
@@ -31,6 +32,35 @@ describe('documentHelper', () => {
     });
   });
 
+  describe('#isSafeHttpLink', () => {
+    it('returns true for http and https URLs', () => {
+      expect(isSafeHttpLink('http://example.com')).toBe(true);
+      expect(isSafeHttpLink('https://example.com/path?q=1#x')).toBe(true);
+      expect(isSafeHttpLink('HTTPS://EXAMPLE.COM')).toBe(true);
+    });
+
+    /* eslint-disable no-script-url */
+    it('returns false for javascript: and other dangerous schemes', () => {
+      expect(isSafeHttpLink('javascript:alert(1)')).toBe(false);
+      expect(isSafeHttpLink('JavaScript:alert(1)')).toBe(false);
+      expect(isSafeHttpLink('data:text/html,<script>alert(1)</script>')).toBe(
+        false
+      );
+      expect(isSafeHttpLink('vbscript:msgbox(1)')).toBe(false);
+      expect(isSafeHttpLink('file:///etc/passwd')).toBe(false);
+      expect(isSafeHttpLink('ftp://files.example.com/doc.pdf')).toBe(false);
+    });
+    /* eslint-enable no-script-url */
+
+    it('returns false for invalid or empty values', () => {
+      expect(isSafeHttpLink('')).toBe(false);
+      expect(isSafeHttpLink(null)).toBe(false);
+      expect(isSafeHttpLink(undefined)).toBe(false);
+      expect(isSafeHttpLink('not a url')).toBe(false);
+      expect(isSafeHttpLink('//example.com')).toBe(false);
+    });
+  });
+
   describe('#formatDocumentLink', () => {
     describe('PDF documents', () => {
       it('removes PDF: prefix from PDF documents', () => {
@@ -78,32 +108,30 @@ describe('documentHelper', () => {
     });
 
     describe('Regular URLs', () => {
-      it('returns regular URLs unchanged', () => {
-        expect(formatDocumentLink('https://example.com')).toBe(
-          'https://example.com'
-        );
+      it('removes http(s) and www prefixes for compact display', () => {
+        expect(formatDocumentLink('https://example.com')).toBe('example.com');
         expect(formatDocumentLink('http://docs.example.com/api')).toBe(
-          'http://docs.example.com/api'
+          'docs.example.com/api'
         );
-        expect(formatDocumentLink('https://github.com/user/repo')).toBe(
-          'https://github.com/user/repo'
+        expect(formatDocumentLink('https://www.github.com/user/repo')).toBe(
+          'github.com/user/repo'
         );
       });
 
       it('handles URLs with query parameters', () => {
         expect(formatDocumentLink('https://example.com?param=value')).toBe(
-          'https://example.com?param=value'
+          'example.com?param=value'
         );
         expect(
           formatDocumentLink(
             'https://api.example.com/docs?version=v1&format=json'
           )
-        ).toBe('https://api.example.com/docs?version=v1&format=json');
+        ).toBe('api.example.com/docs?version=v1&format=json');
       });
 
       it('handles URLs with fragments', () => {
         expect(formatDocumentLink('https://example.com/docs#section1')).toBe(
-          'https://example.com/docs#section1'
+          'example.com/docs#section1'
         );
       });
     });
diff --git a/db/migrate/20260429043000_add_sync_stats_index_to_captain_documents.rb b/db/migrate/20260429043000_add_sync_stats_index_to_captain_documents.rb
new file mode 100644
index 000000000..9c2cdb66c
--- /dev/null
+++ b/db/migrate/20260429043000_add_sync_stats_index_to_captain_documents.rb
@@ -0,0 +1,12 @@
+class AddSyncStatsIndexToCaptainDocuments < ActiveRecord::Migration[7.0]
+  def up
+    add_index :captain_documents,
+              [:account_id, :assistant_id, :sync_status, :last_synced_at],
+              name: 'idx_captain_documents_on_account_assistant_sync_stats',
+              if_not_exists: true
+  end
+
+  def down
+    remove_index :captain_documents, name: 'idx_captain_documents_on_account_assistant_sync_stats', if_exists: true
+  end
+end
diff --git a/db/schema.rb b/db/schema.rb
index 9f8d76b78..1948330e6 100644
--- a/db/schema.rb
+++ b/db/schema.rb
@@ -381,11 +381,12 @@ ActiveRecord::Schema[7.1].define(version: 2026_05_07_000000) do
     t.integer "sync_status"
     t.datetime "last_synced_at"
     t.datetime "last_sync_attempted_at"
+    t.index ["account_id", "assistant_id", "sync_status", "last_synced_at"], name: "idx_captain_documents_on_account_assistant_sync_stats"
+    t.index ["account_id", "sync_status"], name: "index_captain_documents_on_account_id_and_sync_status"
     t.index ["account_id"], name: "index_captain_documents_on_account_id"
     t.index ["assistant_id", "external_link"], name: "index_captain_documents_on_assistant_id_and_external_link", unique: true
     t.index ["assistant_id"], name: "index_captain_documents_on_assistant_id"
     t.index ["status"], name: "index_captain_documents_on_status"
-    t.index ["account_id", "sync_status"], name: "index_captain_documents_on_account_id_and_sync_status"
   end
 
   create_table "captain_inboxes", force: :cascade do |t|
diff --git a/enterprise/app/controllers/api/v1/accounts/captain/bulk_actions_controller.rb b/enterprise/app/controllers/api/v1/accounts/captain/bulk_actions_controller.rb
index bc1ebaf9e..b9a6bbcc5 100644
--- a/enterprise/app/controllers/api/v1/accounts/captain/bulk_actions_controller.rb
+++ b/enterprise/app/controllers/api/v1/accounts/captain/bulk_actions_controller.rb
@@ -77,7 +77,12 @@ class Api::V1::Accounts::Captain::BulkActionsController < Api::V1::Accounts::Bas
       next unless document.available?
       next if document.sync_in_progress?
 
-      document.update!(sync_status: :syncing, last_sync_attempted_at: Time.current)
+      document.update!(
+        sync_status: :syncing,
+        sync_step: nil,
+        last_sync_error_code: nil,
+        last_sync_attempted_at: Time.current
+      )
       Captain::Documents::PerformSyncJob.perform_later(document)
       synced_document_ids << document.id
     end
diff --git a/enterprise/app/controllers/api/v1/accounts/captain/documents_controller.rb b/enterprise/app/controllers/api/v1/accounts/captain/documents_controller.rb
index 973559743..23f410499 100644
--- a/enterprise/app/controllers/api/v1/accounts/captain/documents_controller.rb
+++ b/enterprise/app/controllers/api/v1/accounts/captain/documents_controller.rb
@@ -11,8 +11,13 @@ class Api::V1::Accounts::Captain::DocumentsController < Api::V1::Accounts::BaseC
   def index
     base_query = @documents
     base_query = base_query.where(assistant_id: permitted_params[:assistant_id]) if permitted_params[:assistant_id].present?
+    base_query = apply_source_filter(base_query, permitted_params[:source])
+    base_query = apply_filter(base_query, permitted_params[:filter])
+    base_query = apply_search(base_query, permitted_params[:search_key])
+    base_query = apply_sort(base_query, permitted_params[:sort])
 
     @documents_count = base_query.count
+    @sync_interval_hours = current_sync_interval&.in_hours&.to_i
     @documents = base_query.page(@current_page).per(RESULTS_PER_PAGE)
   end
 
@@ -34,7 +39,12 @@ class Api::V1::Accounts::Captain::DocumentsController < Api::V1::Accounts::BaseC
     return render_could_not_create_error(I18n.t('captain.documents.sync_only_available_documents')) unless @document.available?
     return render_could_not_create_error(I18n.t('captain.documents.sync_already_in_progress')) if @document.sync_in_progress?
 
-    @document.update!(sync_status: :syncing, last_sync_attempted_at: Time.current)
+    @document.update!(
+      sync_status: :syncing,
+      sync_step: nil,
+      last_sync_error_code: nil,
+      last_sync_attempted_at: Time.current
+    )
     Captain::Documents::PerformSyncJob.perform_later(@document)
     head :accepted
   end
@@ -47,7 +57,7 @@ class Api::V1::Accounts::Captain::DocumentsController < Api::V1::Accounts::BaseC
   private
 
   def set_documents
-    @documents = Current.account.captain_documents.includes(:assistant).ordered
+    @documents = Current.account.captain_documents.with_attached_pdf_file.includes(:assistant)
   end
 
   def set_document
@@ -63,7 +73,58 @@ class Api::V1::Accounts::Captain::DocumentsController < Api::V1::Accounts::BaseC
   end
 
   def permitted_params
-    params.permit(:assistant_id, :page, :id, :account_id)
+    params.permit(:assistant_id, :page, :id, :account_id, :filter, :source, :sort, :search_key)
+  end
+
+  def apply_source_filter(scope, source)
+    case source
+    when 'web' then scope.syncable
+    when 'pdf' then scope.pdf_documents
+    else scope
+    end
+  end
+
+  def apply_filter(scope, filter)
+    case filter
+    when 'stale' then stale_documents(scope.syncable)
+    when 'synced' then up_to_date_documents(scope.syncable)
+    when 'syncing' then scope.syncable.sync_in_progress
+    when 'failed' then scope.syncable.sync_failed
+    else scope
+    end
+  end
+
+  def apply_search(scope, search_key)
+    return scope if search_key.blank?
+
+    query = "%#{ActiveRecord::Base.sanitize_sql_like(search_key)}%"
+    scope.where('captain_documents.name ILIKE :query OR captain_documents.external_link ILIKE :query', query: query)
+  end
+
+  def apply_sort(scope, sort)
+    case sort
+    when 'recently_created' then scope.order(created_at: :desc)
+    else scope.order(updated_at: :desc)
+    end
+  end
+
+  def stale_documents(scope)
+    return scope.none unless current_sync_interval
+
+    scope.sync_synced.where(Captain::Document.arel_table[:last_synced_at].lt(current_sync_interval.ago))
+  end
+
+  def up_to_date_documents(scope)
+    documents = scope.sync_synced
+    return documents unless current_sync_interval
+
+    documents.where(Captain::Document.arel_table[:last_synced_at].gteq(current_sync_interval.ago))
+  end
+
+  def current_sync_interval
+    return @current_sync_interval if defined?(@current_sync_interval)
+
+    @current_sync_interval = Current.account.captain_document_sync_interval
   end
 
   def document_params
diff --git a/enterprise/app/jobs/captain/documents/perform_sync_job.rb b/enterprise/app/jobs/captain/documents/perform_sync_job.rb
index 100d72eeb..eaef7d64d 100644
--- a/enterprise/app/jobs/captain/documents/perform_sync_job.rb
+++ b/enterprise/app/jobs/captain/documents/perform_sync_job.rb
@@ -20,13 +20,13 @@ class Captain::Documents::PerformSyncJob < MutexApplicationJob
                                           exception_class: error.class.name)
   end
 
-  # Permanent errors (404, 403, empty content) — no point retrying, discard immediately.
+  # Permanent errors (404, 403, empty content) - no point retrying, discard immediately.
   # Document is already marked failed by SyncService before the exception reaches here.
   discard_on(Captain::Documents::SyncService::PermanentSyncError)
 
-  # TransientSyncError is raised by SyncService when the customer's site is unreachable —
+  # TransientSyncError is raised by SyncService when the customer's site is unreachable -
   # timeouts, TLS errors, 5xx, connection drops. Four attempts with backoff gives the site
-  # a chance to recover before we give up.
+  # a chance to recover before we mark the document failed.
   #
   # The exhaustion block absorbs the exception so it doesn't propagate to Sentry —
   # site flakiness isn't an application bug.
@@ -36,6 +36,7 @@ class Captain::Documents::PerformSyncJob < MutexApplicationJob
     attempts: 4
   ) do |job, error|
     document = job.arguments.first
+    job.send(:mark_sync_failed, document, error.message)
     job.send(:log_sync_outcome, document, result: :transient_retry_exhausted, error_code: error.message)
   end
 
@@ -47,7 +48,7 @@ class Captain::Documents::PerformSyncJob < MutexApplicationJob
     return if document.pdf_document?
 
     with_lock(lock_key(document), LOCK_TIMEOUT) do
-      document.update!(sync_status: :syncing, last_sync_attempted_at: Time.current)
+      mark_sync_started(document)
       result = Captain::Documents::SyncService.new(document.reload).perform
       log_sync_outcome(document, result: result, duration_ms: duration_ms_since(start_time))
     end
@@ -78,13 +79,26 @@ class Captain::Documents::PerformSyncJob < MutexApplicationJob
     raise error
   end
 
-  def handle_unexpected_failure(document, error, start_time)
+  def mark_sync_failed(document, error_code)
     document.update!(
       sync_status: :failed,
       sync_step: nil,
-      last_sync_error_code: 'sync_error',
+      last_sync_error_code: error_code,
       last_sync_attempted_at: Time.current
     )
+  end
+
+  def mark_sync_started(document)
+    document.update!(
+      sync_status: :syncing,
+      sync_step: nil,
+      last_sync_error_code: nil,
+      last_sync_attempted_at: Time.current
+    )
+  end
+
+  def handle_unexpected_failure(document, error, start_time)
+    mark_sync_failed(document, 'sync_error')
     log_sync_outcome(document, result: :unexpected_failure, error_code: 'sync_error',
                                exception_class: error.class.name,
                                duration_ms: duration_ms_since(start_time))
diff --git a/enterprise/app/jobs/captain/documents/schedule_syncs_job.rb b/enterprise/app/jobs/captain/documents/schedule_syncs_job.rb
index 4103fd271..393a307db 100644
--- a/enterprise/app/jobs/captain/documents/schedule_syncs_job.rb
+++ b/enterprise/app/jobs/captain/documents/schedule_syncs_job.rb
@@ -82,7 +82,7 @@ class Captain::Documents::ScheduleSyncsJob < ApplicationJob
   end
 
   def reserve_sync_slot(document)
-    document.update!(sync_status: :syncing, last_sync_attempted_at: Time.current)
+    mark_sync_started(document)
     true
   rescue ActiveRecord::RecordInvalid => e
     log_document_skip(document, e)
@@ -112,4 +112,13 @@ class Captain::Documents::ScheduleSyncsJob < ApplicationJob
 
     Rails.logger.info("[Captain::Documents::ScheduleSyncsJob] #{payload.to_json}")
   end
+
+  def mark_sync_started(document)
+    document.update!(
+      sync_status: :syncing,
+      sync_step: nil,
+      last_sync_error_code: nil,
+      last_sync_attempted_at: Time.current
+    )
+  end
 end
diff --git a/enterprise/app/models/captain/document.rb b/enterprise/app/models/captain/document.rb
index 7509286a1..0fe14813b 100644
--- a/enterprise/app/models/captain/document.rb
+++ b/enterprise/app/models/captain/document.rb
@@ -4,8 +4,10 @@
 #
 #  id                     :bigint           not null, primary key
 #  content                :text
+#  content_fingerprint    :string
 #  external_link          :string           not null
 #  last_sync_attempted_at :datetime
+#  last_sync_error_code   :string
 #  last_synced_at         :datetime
 #  metadata               :jsonb
 #  name                   :string
@@ -18,6 +20,7 @@
 #
 # Indexes
 #
+#  idx_captain_documents_on_account_assistant_sync_stats      (account_id,assistant_id,sync_status,last_synced_at)
 #  index_captain_documents_on_account_id                      (account_id)
 #  index_captain_documents_on_account_id_and_sync_status      (account_id,sync_status)
 #  index_captain_documents_on_assistant_id                    (assistant_id)
@@ -62,6 +65,14 @@ class Captain::Document < ApplicationRecord
   scope :for_account, ->(account_id) { where(account_id: account_id) }
   scope :for_assistant, ->(assistant_id) { where(assistant_id: assistant_id) }
   scope :syncable, -> { where("external_link NOT LIKE 'PDF:%' AND external_link NOT LIKE '%.pdf'") }
+  scope :pdf_documents, -> { where("external_link LIKE 'PDF:%' OR external_link LIKE '%.pdf'") }
+  scope :sync_in_progress, -> { sync_syncing.where(arel_table[:last_sync_attempted_at].gteq(SYNC_STALE_TIMEOUT.ago)) }
+  scope :stale, lambda { |stale_before|
+    sync_failed.or(sync_synced.where(arel_table[:last_synced_at].lt(stale_before)))
+  }
+  scope :synced_since, lambda { |time|
+    sync_synced.where(arel_table[:last_synced_at].gteq(time))
+  }
 
   def pdf_document?
     return true if pdf_file.attached? && pdf_file.blob.content_type == 'application/pdf'
diff --git a/enterprise/app/policies/captain/assistant_policy.rb b/enterprise/app/policies/captain/assistant_policy.rb
index a3cc19b16..bbde3ffb0 100644
--- a/enterprise/app/policies/captain/assistant_policy.rb
+++ b/enterprise/app/policies/captain/assistant_policy.rb
@@ -7,6 +7,10 @@ class Captain::AssistantPolicy < ApplicationPolicy
     true
   end
 
+  def stats?
+    true
+  end
+
   def tools?
     @account_user.administrator?
   end
diff --git a/enterprise/app/services/captain/documents/sync_service.rb b/enterprise/app/services/captain/documents/sync_service.rb
index f1a67c781..eca28dbab 100644
--- a/enterprise/app/services/captain/documents/sync_service.rb
+++ b/enterprise/app/services/captain/documents/sync_service.rb
@@ -15,10 +15,7 @@ class Captain::Documents::SyncService
     @document.update!(sync_step: 'fetching')
     result = Captain::Documents::SinglePageFetcher.new(@document.external_link).fetch
 
-    unless result.success
-      mark_failed(result.error_code)
-      raise_for_error_code(result.error_code)
-    end
+    handle_fetch_error(result.error_code) unless result.success
 
     @document.update!(sync_step: 'comparing')
     new_fingerprint = compute_fingerprint(result.content)
@@ -75,8 +72,11 @@ class Captain::Documents::SyncService
     )
   end
 
-  def raise_for_error_code(error_code)
-    raise PermanentSyncError, error_code if PERMANENT_ERROR_CODES.include?(error_code)
+  def handle_fetch_error(error_code)
+    if PERMANENT_ERROR_CODES.include?(error_code)
+      mark_failed(error_code)
+      raise PermanentSyncError, error_code
+    end
 
     raise TransientSyncError, error_code
   end
diff --git a/enterprise/app/views/api/v1/accounts/captain/documents/index.json.jbuilder b/enterprise/app/views/api/v1/accounts/captain/documents/index.json.jbuilder
index 5b8c726ea..e6fb5fd0b 100644
--- a/enterprise/app/views/api/v1/accounts/captain/documents/index.json.jbuilder
+++ b/enterprise/app/views/api/v1/accounts/captain/documents/index.json.jbuilder
@@ -7,4 +7,5 @@ end
 json.meta do
   json.total_count @documents_count
   json.page @current_page
+  json.sync_interval_hours @sync_interval_hours if @sync_interval_hours.present?
 end
diff --git a/enterprise/app/views/api/v1/models/captain/_document.json.jbuilder b/enterprise/app/views/api/v1/models/captain/_document.json.jbuilder
index 62710bc64..56260f675 100644
--- a/enterprise/app/views/api/v1/models/captain/_document.json.jbuilder
+++ b/enterprise/app/views/api/v1/models/captain/_document.json.jbuilder
@@ -8,10 +8,12 @@ json.created_at resource.created_at.to_i
 json.external_link resource.external_link
 json.display_url resource.display_url
 json.file_size resource.file_size
+json.pdf_document resource.pdf_document?
 json.id resource.id
 json.name resource.name
 json.status resource.status
 json.sync_status resource.sync_status
+json.sync_in_progress resource.sync_in_progress?
 json.last_synced_at resource.last_synced_at&.to_i
 json.last_sync_attempted_at resource.last_sync_attempted_at&.to_i
 json.last_sync_error_code resource.last_sync_error_code

From 086aa36ffedc7a1d6269f10991a5a9d9bdc9dd55 Mon Sep 17 00:00:00 2001
From: Sojan Jose <sojan@pepalo.com>
Date: Mon, 11 May 2026 23:15:25 +0530
Subject: [PATCH 13/20] feat(companies): add notes and history to company
 details (#14401)

---
 app/javascript/dashboard/api/companies.js     |   8 ++
 .../Companies/CompaniesDetailsLayout.vue      |  18 ++-
 .../CompanyDetail/CompanyContactsSidebar.vue  |   6 +-
 .../CompanyDetail/CompanyHistorySidebar.vue   |  67 ++++++++++
 .../CompanyDetail/CompanyNotesSidebar.vue     | 114 ++++++++++++++++++
 .../dashboard/i18n/locale/en/companies.json   |  10 +-
 .../companies/pages/CompanyDetailView.vue     |  69 +++++++++++
 .../dashboard/store/storeFactory.js           |   7 +-
 app/javascript/dashboard/stores/companies.js  |  85 +++++++++++++
 config/routes.rb                              |   2 +
 .../v1/accounts/companies/base_controller.rb  |  24 ++++
 .../accounts/companies/contacts_controller.rb |  22 +---
 .../companies/conversations_controller.rb     |  15 +++
 .../v1/accounts/companies/notes_controller.rb |  11 ++
 .../conversations/index.json.jbuilder         |   5 +
 .../companies/notes/index.json.jbuilder       |   8 ++
 16 files changed, 440 insertions(+), 31 deletions(-)
 create mode 100644 app/javascript/dashboard/components-next/Companies/CompanyDetail/CompanyHistorySidebar.vue
 create mode 100644 app/javascript/dashboard/components-next/Companies/CompanyDetail/CompanyNotesSidebar.vue
 create mode 100644 enterprise/app/controllers/api/v1/accounts/companies/base_controller.rb
 create mode 100644 enterprise/app/controllers/api/v1/accounts/companies/conversations_controller.rb
 create mode 100644 enterprise/app/controllers/api/v1/accounts/companies/notes_controller.rb
 create mode 100644 enterprise/app/views/api/v1/accounts/companies/conversations/index.json.jbuilder
 create mode 100644 enterprise/app/views/api/v1/accounts/companies/notes/index.json.jbuilder

diff --git a/app/javascript/dashboard/api/companies.js b/app/javascript/dashboard/api/companies.js
index b98a59983..a45b21d68 100644
--- a/app/javascript/dashboard/api/companies.js
+++ b/app/javascript/dashboard/api/companies.js
@@ -28,6 +28,14 @@ class CompanyAPI extends ApiClient {
     return axios.get(`${this.url}/${id}/contacts?${buildParams({ page })}`);
   }
 
+  listNotes(id) {
+    return axios.get(`${this.url}/${id}/notes`);
+  }
+
+  listConversations(id) {
+    return axios.get(`${this.url}/${id}/conversations`);
+  }
+
   searchContacts(id, query = '', page = 1) {
     const requestURL = `${this.url}/${id}/contacts/search?${buildParams({ q: query, page })}`;
     return axios.get(requestURL);
diff --git a/app/javascript/dashboard/components-next/Companies/CompaniesDetailsLayout.vue b/app/javascript/dashboard/components-next/Companies/CompaniesDetailsLayout.vue
index 7736ccffd..14b5e62d9 100644
--- a/app/javascript/dashboard/components-next/Companies/CompaniesDetailsLayout.vue
+++ b/app/javascript/dashboard/components-next/Companies/CompaniesDetailsLayout.vue
@@ -56,9 +56,14 @@ const closeMobileSidebar = () => {
 
     <div
       v-if="slots.sidebar"
-      class="hidden lg:block overflow-y-auto justify-end min-w-52 w-full py-6 max-w-md border-l border-n-weak bg-n-solid-2"
+      class="hidden lg:flex flex-col min-w-52 w-full max-w-md border-l border-n-weak bg-n-solid-2"
     >
-      <slot name="sidebar" />
+      <div class="shrink-0">
+        <slot name="sidebarHeader" />
+      </div>
+      <div class="flex-1 overflow-y-auto pb-6 pt-3">
+        <slot name="sidebar" />
+      </div>
     </div>
 
     <div
@@ -105,9 +110,14 @@ const closeMobileSidebar = () => {
         <div
           v-if="isSidebarOpen"
           id="details-sidebar-content"
-          class="order-2 w-[85%] sm:w-[50%] bg-n-solid-2 ltr:border-l rtl:border-r border-n-weak overflow-y-auto py-6 shadow-lg"
+          class="order-2 w-[85%] sm:w-[50%] flex flex-col bg-n-solid-2 ltr:border-l rtl:border-r border-n-weak shadow-lg"
         >
-          <slot name="sidebar" />
+          <div class="shrink-0">
+            <slot name="sidebarHeader" />
+          </div>
+          <div class="flex-1 overflow-y-auto pb-6 pt-3">
+            <slot name="sidebar" />
+          </div>
         </div>
       </Transition>
     </div>
diff --git a/app/javascript/dashboard/components-next/Companies/CompanyDetail/CompanyContactsSidebar.vue b/app/javascript/dashboard/components-next/Companies/CompanyDetail/CompanyContactsSidebar.vue
index 7e0d162dc..ce2a68da1 100644
--- a/app/javascript/dashboard/components-next/Companies/CompanyDetail/CompanyContactsSidebar.vue
+++ b/app/javascript/dashboard/components-next/Companies/CompanyDetail/CompanyContactsSidebar.vue
@@ -169,7 +169,7 @@ const handleContactSelect = contactId => {
 </script>
 
 <template>
-  <div class="flex flex-col gap-6 px-6 pb-8 pt-1">
+  <div class="flex flex-col gap-6 px-6 pb-8">
     <div v-if="!selectedContact" class="flex flex-col gap-4">
       <div class="flex flex-col gap-2">
         <label class="text-base text-n-slate-12">
@@ -288,7 +288,7 @@ const handleContactSelect = contactId => {
 
       <div
         v-else-if="!hasContacts"
-        class="py-8 text-sm text-center rounded-xl border border-dashed border-n-weak text-n-slate-11"
+        class="py-8 px-4 text-sm text-center rounded-xl border border-dashed border-n-strong text-n-slate-11"
       >
         {{ t('COMPANIES.DETAIL.CONTACTS.EMPTY') }}
       </div>
@@ -346,7 +346,7 @@ const handleContactSelect = contactId => {
         :current-page="currentPage"
         :total-items="totalContacts"
         :items-per-page="15"
-        class="!px-0 before:hidden"
+        class="!px-0 before:hidden bg-transparent"
         @update:current-page="emit('update:currentPage', $event)"
       />
     </div>
diff --git a/app/javascript/dashboard/components-next/Companies/CompanyDetail/CompanyHistorySidebar.vue b/app/javascript/dashboard/components-next/Companies/CompanyDetail/CompanyHistorySidebar.vue
new file mode 100644
index 000000000..c82ed71fb
--- /dev/null
+++ b/app/javascript/dashboard/components-next/Companies/CompanyDetail/CompanyHistorySidebar.vue
@@ -0,0 +1,67 @@
+<script setup>
+import { computed } from 'vue';
+import { useI18n } from 'vue-i18n';
+import { useMapGetter } from 'dashboard/composables/store';
+
+import ConversationCard from 'dashboard/components-next/Conversation/ConversationCard/ConversationCard.vue';
+import Spinner from 'dashboard/components-next/spinner/Spinner.vue';
+
+defineProps({
+  conversations: {
+    type: Array,
+    default: () => [],
+  },
+  isLoading: {
+    type: Boolean,
+    default: false,
+  },
+});
+
+const { t } = useI18n();
+const contactsById = useMapGetter('contacts/getContactById');
+const stateInbox = useMapGetter('inboxes/getInboxById');
+const accountLabels = useMapGetter('labels/getLabels');
+
+const accountLabelsValue = computed(() => accountLabels.value);
+const conversationContact = conversation => {
+  const sender = conversation.meta?.sender || {};
+  const contact = contactsById.value(sender.id);
+  return contact.id ? contact : sender;
+};
+const conversationInbox = conversation =>
+  stateInbox.value(conversation.inboxId) || {
+    name: '',
+    channelType: conversation.meta?.channel,
+  };
+</script>
+
+<template>
+  <div
+    v-if="isLoading"
+    class="flex items-center justify-center py-10 text-n-slate-11"
+  >
+    <Spinner />
+  </div>
+
+  <div
+    v-else-if="conversations.length > 0"
+    class="px-6 divide-y divide-n-strong [&>*:hover]:!border-y-transparent [&>*:hover+*]:!border-t-transparent"
+  >
+    <ConversationCard
+      v-for="conversation in conversations"
+      :key="conversation.id"
+      :conversation="conversation"
+      :contact="conversationContact(conversation)"
+      :state-inbox="conversationInbox(conversation)"
+      :account-labels="accountLabelsValue"
+      class="rounded-none hover:rounded-xl hover:bg-n-alpha-1 dark:hover:bg-n-alpha-3"
+    />
+  </div>
+
+  <p
+    v-else
+    class="py-8 px-4 mx-6 text-sm text-center rounded-xl border border-dashed border-n-strong text-n-slate-11"
+  >
+    {{ t('COMPANIES.DETAIL.HISTORY.EMPTY') }}
+  </p>
+</template>
diff --git a/app/javascript/dashboard/components-next/Companies/CompanyDetail/CompanyNotesSidebar.vue b/app/javascript/dashboard/components-next/Companies/CompanyDetail/CompanyNotesSidebar.vue
new file mode 100644
index 000000000..234e0cd12
--- /dev/null
+++ b/app/javascript/dashboard/components-next/Companies/CompanyDetail/CompanyNotesSidebar.vue
@@ -0,0 +1,114 @@
+<script setup>
+import { computed } from 'vue';
+import { useI18n } from 'vue-i18n';
+import { useRoute, useRouter } from 'vue-router';
+import { useMapGetter } from 'dashboard/composables/store';
+import { dynamicTime } from 'shared/helpers/timeHelper';
+import { useMessageFormatter } from 'shared/composables/useMessageFormatter';
+
+import Avatar from 'dashboard/components-next/avatar/Avatar.vue';
+import Spinner from 'dashboard/components-next/spinner/Spinner.vue';
+
+const props = defineProps({
+  notes: {
+    type: Array,
+    default: () => [],
+  },
+  isLoading: {
+    type: Boolean,
+    default: false,
+  },
+});
+
+const route = useRoute();
+const router = useRouter();
+const { t } = useI18n();
+const { formatMessage } = useMessageFormatter();
+const currentUser = useMapGetter('getCurrentUser');
+
+const hasNotes = computed(() => props.notes.length > 0);
+
+const contactName = contact =>
+  contact?.name || t('COMPANIES.DETAIL.CONTACTS.UNNAMED_CONTACT');
+
+const getWrittenBy = note => {
+  const isCurrentUser = note?.user?.id === currentUser.value.id;
+  return isCurrentUser
+    ? t('CONTACTS_LAYOUT.SIDEBAR.NOTES.YOU')
+    : note?.user?.name || 'Bot';
+};
+
+const openContact = contactId => {
+  router.push({
+    name: 'contacts_edit',
+    params: {
+      accountId: route.params.accountId,
+      contactId,
+    },
+  });
+};
+</script>
+
+<template>
+  <div v-if="hasNotes" class="flex flex-col px-6">
+    <div class="flex flex-col divide-y divide-n-strong">
+      <div
+        v-for="note in notes"
+        :key="note.id"
+        class="flex flex-col gap-2 py-4 group/note"
+      >
+        <div class="flex items-center gap-1.5 min-w-0">
+          <Avatar
+            :name="contactName(note.contact)"
+            :src="note.contact?.thumbnail"
+            :size="16"
+            rounded-full
+            hide-offline-status
+          />
+          <div
+            class="flex items-center justify-between min-w-0 gap-1 w-full text-sm text-n-slate-11"
+          >
+            <button
+              type="button"
+              class="min-w-0 font-medium truncate text-start text-n-slate-12 hover:text-n-blue-11 p-0"
+              @click="openContact(note.contact.id)"
+            >
+              {{ contactName(note.contact) }}
+            </button>
+            <div class="min-w-0 truncate">
+              <span
+                class="inline-flex items-center gap-1 text-sm text-n-slate-10"
+              >
+                <span class="font-medium text-n-slate-11">
+                  {{ getWrittenBy(note) }}
+                </span>
+                {{ t('CONTACTS_LAYOUT.SIDEBAR.NOTES.WROTE') }}
+                <span class="font-medium text-n-slate-11">
+                  {{ dynamicTime(note.createdAt) }}
+                </span>
+              </span>
+            </div>
+          </div>
+        </div>
+        <p
+          v-dompurify-html="formatMessage(note.content || '')"
+          class="mb-0 prose-sm prose-p:text-sm prose-p:leading-relaxed prose-p:mb-1 prose-p:mt-0 prose-ul:mb-1 prose-ul:mt-0 text-n-slate-12"
+        />
+      </div>
+    </div>
+  </div>
+
+  <div
+    v-else-if="isLoading"
+    class="flex items-center justify-center py-10 text-n-slate-11"
+  >
+    <Spinner />
+  </div>
+
+  <p
+    v-else
+    class="py-8 mx-6 px-4 text-sm text-center rounded-xl border border-dashed border-n-strong text-n-slate-11"
+  >
+    {{ t('COMPANIES.DETAIL.NOTES.EMPTY') }}
+  </p>
+</template>
diff --git a/app/javascript/dashboard/i18n/locale/en/companies.json b/app/javascript/dashboard/i18n/locale/en/companies.json
index 1fe8bc169..534205038 100644
--- a/app/javascript/dashboard/i18n/locale/en/companies.json
+++ b/app/javascript/dashboard/i18n/locale/en/companies.json
@@ -44,9 +44,17 @@
       "SIDEBAR": {
         "TABS": {
           "ATTRIBUTES": "Attributes",
-          "CONTACTS": "Contacts"
+          "CONTACTS": "Contacts",
+          "HISTORY": "History",
+          "NOTES": "Notes"
         }
       },
+      "HISTORY": {
+        "EMPTY": "No conversations found for this company's contacts yet."
+      },
+      "NOTES": {
+        "EMPTY": "No notes found for this company's contacts yet."
+      },
       "ATTRIBUTES": {
         "SEARCH_PLACEHOLDER": "Search attributes...",
         "EMPTY_STATE": "There are no company custom attributes configured yet.",
diff --git a/app/javascript/dashboard/routes/dashboard/companies/pages/CompanyDetailView.vue b/app/javascript/dashboard/routes/dashboard/companies/pages/CompanyDetailView.vue
index 313df49d3..86627f952 100644
--- a/app/javascript/dashboard/routes/dashboard/companies/pages/CompanyDetailView.vue
+++ b/app/javascript/dashboard/routes/dashboard/companies/pages/CompanyDetailView.vue
@@ -8,7 +8,10 @@ import Policy from 'dashboard/components/policy.vue';
 import Button from 'dashboard/components-next/button/Button.vue';
 import CompaniesDetailsLayout from 'dashboard/components-next/Companies/CompaniesDetailsLayout.vue';
 import Spinner from 'dashboard/components-next/spinner/Spinner.vue';
+import TabBar from 'dashboard/components-next/tabbar/TabBar.vue';
 import CompanyContactsSidebar from 'dashboard/components-next/Companies/CompanyDetail/CompanyContactsSidebar.vue';
+import CompanyHistorySidebar from 'dashboard/components-next/Companies/CompanyDetail/CompanyHistorySidebar.vue';
+import CompanyNotesSidebar from 'dashboard/components-next/Companies/CompanyDetail/CompanyNotesSidebar.vue';
 import CompanyProfileCard from 'dashboard/components-next/Companies/CompanyDetail/CompanyProfileCard.vue';
 import ConfirmCompanyDeleteDialog from 'dashboard/components-next/Companies/CompanyDetail/ConfirmCompanyDeleteDialog.vue';
 import { useCompaniesStore } from 'dashboard/stores/companies';
@@ -20,11 +23,16 @@ const { t } = useI18n();
 
 const confirmDeleteDialogRef = ref(null);
 const selectedCandidate = ref(null);
+const activeSidebarTab = ref('history');
 
 const companyId = computed(() => Number(route.params.companyId));
 const company = computed(() => companiesStore.getRecord(companyId.value));
 const companyContacts = computed(() => companiesStore.companyContacts);
 const companyContactsMeta = computed(() => companiesStore.companyContactsMeta);
+const companyConversations = computed(
+  () => companiesStore.companyConversations || []
+);
+const companyNotes = computed(() => companiesStore.companyNotes || []);
 const contactSearchResults = computed(
   () => companiesStore.contactSearchResults
 );
@@ -32,6 +40,10 @@ const uiFlags = computed(() => companiesStore.getUIFlags);
 
 const isFetchingCompany = computed(() => uiFlags.value.fetchingItem);
 const isFetchingContacts = computed(() => uiFlags.value.fetchingContacts);
+const isFetchingConversations = computed(
+  () => uiFlags.value.fetchingConversations
+);
+const isFetchingNotes = computed(() => uiFlags.value.fetchingNotes);
 const isSearchingContacts = computed(() => uiFlags.value.searchingContacts);
 const isManagingContacts = computed(
   () => uiFlags.value.creatingContact || uiFlags.value.removingContact
@@ -50,6 +62,27 @@ const breadcrumbItems = computed(() => [
     : []),
 ]);
 
+const SIDEBAR_TABS_OPTIONS = [
+  { key: 'HISTORY', value: 'history' },
+  { key: 'NOTES', value: 'notes' },
+  { key: 'CONTACTS', value: 'contacts' },
+];
+
+const sidebarTabs = computed(() =>
+  SIDEBAR_TABS_OPTIONS.map(tab => ({
+    label: {
+      notes: t('COMPANIES.DETAIL.SIDEBAR.TABS.NOTES'),
+      history: t('COMPANIES.DETAIL.SIDEBAR.TABS.HISTORY'),
+      contacts: `${t('COMPANIES.DETAIL.SIDEBAR.TABS.CONTACTS')} (${Number(companyContactsMeta.value.totalCount || 0)})`,
+    }[tab.value],
+    value: tab.value,
+  }))
+);
+
+const activeSidebarTabIndex = computed(() =>
+  SIDEBAR_TABS_OPTIONS.findIndex(tab => tab.value === activeSidebarTab.value)
+);
+
 const goToCompaniesIndex = () => {
   router.push({
     name: 'companies_dashboard_index',
@@ -79,6 +112,19 @@ const clearSelectedCandidate = () => {
   selectedCandidate.value = null;
 };
 
+const loadSidebarTab = tab => {
+  if (!companyId.value) return;
+  if (tab === 'notes') companiesStore.getCompanyNotes(companyId.value);
+  if (tab === 'history') {
+    companiesStore.getCompanyConversations(companyId.value);
+  }
+};
+
+const handleSidebarTabChange = tab => {
+  activeSidebarTab.value = tab.value;
+  loadSidebarTab(tab.value);
+};
+
 const handleContactSearch = async query => {
   await companiesStore.searchCompanyContactCandidates({
     companyId: companyId.value,
@@ -143,10 +189,12 @@ watch(
   async id => {
     companiesStore.resetCompanyDetailState();
     clearSelectedCandidate();
+    activeSidebarTab.value = 'history';
     if (!id) return;
     await Promise.allSettled([
       companiesStore.show(id),
       companiesStore.getCompanyContacts(id),
+      companiesStore.getCompanyConversations(id),
     ]);
   },
   { immediate: true }
@@ -207,8 +255,29 @@ onBeforeUnmount(() => {
       </Policy>
     </div>
 
+    <template #sidebarHeader>
+      <div class="px-6 pt-6 pb-3">
+        <TabBar
+          :tabs="sidebarTabs"
+          :initial-active-tab="activeSidebarTabIndex"
+          class="w-full [&>button]:w-full bg-n-alpha-black2"
+          @tab-changed="handleSidebarTabChange"
+        />
+      </div>
+    </template>
     <template v-if="hasCompany" #sidebar>
+      <CompanyNotesSidebar
+        v-if="activeSidebarTab === 'notes'"
+        :notes="companyNotes"
+        :is-loading="isFetchingNotes"
+      />
+      <CompanyHistorySidebar
+        v-if="activeSidebarTab === 'history'"
+        :conversations="companyConversations"
+        :is-loading="isFetchingConversations"
+      />
       <CompanyContactsSidebar
+        v-if="activeSidebarTab === 'contacts'"
         :company="company"
         :contacts="companyContacts"
         :meta="companyContactsMeta"
diff --git a/app/javascript/dashboard/store/storeFactory.js b/app/javascript/dashboard/store/storeFactory.js
index a623b4785..9a6d7ec01 100644
--- a/app/javascript/dashboard/store/storeFactory.js
+++ b/app/javascript/dashboard/store/storeFactory.js
@@ -134,10 +134,13 @@ export const createVuexStore = options => {
  * @returns {Function} Pinia store composable
  */
 export const createPiniaStore = options => {
-  const { name, API, actions, getters } = options;
+  const { name, API, actions, getters, state } = options;
 
   return defineStore(name.toLowerCase(), {
-    state: createInitialState,
+    state: () => ({
+      ...createInitialState(),
+      ...(state ? state() : {}),
+    }),
 
     getters: {
       ...createGetters(),
diff --git a/app/javascript/dashboard/stores/companies.js b/app/javascript/dashboard/stores/companies.js
index 1a0d53b41..818e4829f 100644
--- a/app/javascript/dashboard/stores/companies.js
+++ b/app/javascript/dashboard/stores/companies.js
@@ -13,6 +13,8 @@ const createInitialUIFlags = () => ({
   deletingAvatar: false,
   deletingCustomAttributes: false,
   fetchingContacts: false,
+  fetchingConversations: false,
+  fetchingNotes: false,
   searchingContacts: false,
   creatingContact: false,
   removingContact: false,
@@ -67,6 +69,16 @@ export const useCompaniesStore = createStore({
   name: 'companies',
   type: 'pinia',
   API: CompanyAPI,
+  state: () => ({
+    activeCompanyId: null,
+    companyContacts: [],
+    companyContactsMeta: {},
+    companyConversations: [],
+    companyNotes: [],
+    contactSearchResults: [],
+    contactSearchMeta: {},
+    activeContactSearchQuery: '',
+  }),
 
   getters: {
     getCompaniesList: state => state.records,
@@ -270,6 +282,74 @@ export const useCompaniesStore = createStore({
       }
     },
 
+    async getCompanyNotes(companyId) {
+      this.setUIFlag({ fetchingNotes: true });
+      this.ensureActiveCompanyContext(companyId);
+      const activeCompanyId = Number(companyId);
+      const requestToken = (this.companyNotesRequestToken || 0) + 1;
+      this.companyNotesRequestToken = requestToken;
+
+      try {
+        const {
+          data: { payload },
+        } = await CompanyAPI.listNotes(companyId);
+        const notes = camelcaseKeys(payload || [], { deep: true });
+
+        if (
+          this.companyNotesRequestToken !== requestToken ||
+          this.activeCompanyId !== activeCompanyId
+        ) {
+          return notes;
+        }
+
+        this.companyNotes = notes;
+        return notes;
+      } catch (error) {
+        return throwErrorMessage(error);
+      } finally {
+        if (
+          this.companyNotesRequestToken === requestToken &&
+          this.activeCompanyId === activeCompanyId
+        ) {
+          this.setUIFlag({ fetchingNotes: false });
+        }
+      }
+    },
+
+    async getCompanyConversations(companyId) {
+      this.setUIFlag({ fetchingConversations: true });
+      this.ensureActiveCompanyContext(companyId);
+      const activeCompanyId = Number(companyId);
+      const requestToken = (this.companyConversationsRequestToken || 0) + 1;
+      this.companyConversationsRequestToken = requestToken;
+
+      try {
+        const {
+          data: { payload },
+        } = await CompanyAPI.listConversations(companyId);
+        const conversations = camelcaseKeys(payload || [], { deep: true });
+
+        if (
+          this.companyConversationsRequestToken !== requestToken ||
+          this.activeCompanyId !== activeCompanyId
+        ) {
+          return conversations;
+        }
+
+        this.companyConversations = conversations;
+        return conversations;
+      } catch (error) {
+        return throwErrorMessage(error);
+      } finally {
+        if (
+          this.companyConversationsRequestToken === requestToken &&
+          this.activeCompanyId === activeCompanyId
+        ) {
+          this.setUIFlag({ fetchingConversations: false });
+        }
+      }
+    },
+
     async searchCompanyContactCandidates({ companyId, search, page = 1 }) {
       const query = search?.trim() || '';
       if (!query) {
@@ -379,10 +459,15 @@ export const useCompaniesStore = createStore({
         (this.companyDetailRequestToken || 0) + 1;
       this.companyContactsRequestToken =
         (this.companyContactsRequestToken || 0) + 1;
+      this.companyConversationsRequestToken =
+        (this.companyConversationsRequestToken || 0) + 1;
+      this.companyNotesRequestToken = (this.companyNotesRequestToken || 0) + 1;
       this.contactSearchRequestToken =
         (this.contactSearchRequestToken || 0) + 1;
       this.companyContacts = [];
       this.companyContactsMeta = {};
+      this.companyConversations = [];
+      this.companyNotes = [];
       this.contactSearchResults = [];
       this.contactSearchMeta = {};
       this.activeContactSearchQuery = '';
diff --git a/config/routes.rb b/config/routes.rb
index fde0f64dd..ba6a33940 100644
--- a/config/routes.rb
+++ b/config/routes.rb
@@ -187,6 +187,8 @@ Rails.application.routes.draw do
                   get :search
                 end
               end
+              resources :conversations, only: [:index]
+              resources :notes, only: [:index]
             end
           end
           resources :contacts, only: [:index, :show, :update, :create, :destroy] do
diff --git a/enterprise/app/controllers/api/v1/accounts/companies/base_controller.rb b/enterprise/app/controllers/api/v1/accounts/companies/base_controller.rb
new file mode 100644
index 000000000..197a6425c
--- /dev/null
+++ b/enterprise/app/controllers/api/v1/accounts/companies/base_controller.rb
@@ -0,0 +1,24 @@
+class Api::V1::Accounts::Companies::BaseController < Api::V1::Accounts::EnterpriseAccountsController
+  before_action :ensure_companies_enabled!
+  before_action :fetch_company
+
+  private
+
+  def ensure_companies_enabled!
+    return if Current.account.feature_enabled?('companies')
+
+    render json: { error: 'Companies are not enabled for this account' }, status: :forbidden
+  end
+
+  def fetch_company
+    @company = Current.account.companies.find(params[:company_id])
+  end
+
+  def authorize_company_read!
+    authorize(@company, :show?)
+  end
+
+  def authorize_company_update!
+    authorize(@company, :update?)
+  end
+end
diff --git a/enterprise/app/controllers/api/v1/accounts/companies/contacts_controller.rb b/enterprise/app/controllers/api/v1/accounts/companies/contacts_controller.rb
index a67ad907b..07c7b8a99 100644
--- a/enterprise/app/controllers/api/v1/accounts/companies/contacts_controller.rb
+++ b/enterprise/app/controllers/api/v1/accounts/companies/contacts_controller.rb
@@ -1,4 +1,4 @@
-class Api::V1::Accounts::Companies::ContactsController < Api::V1::Accounts::EnterpriseAccountsController
+class Api::V1::Accounts::Companies::ContactsController < Api::V1::Accounts::Companies::BaseController
   RESULTS_PER_PAGE = 15
   CONTACT_SEARCH_QUERY = [
     'contacts.name ILIKE :search',
@@ -7,8 +7,6 @@ class Api::V1::Accounts::Companies::ContactsController < Api::V1::Accounts::Ente
     'contacts.identifier ILIKE :search'
   ].join(' OR ')
 
-  before_action :ensure_companies_enabled!
-  before_action :fetch_company
   before_action :authorize_company_read!, only: [:index, :search]
   before_action :authorize_company_update!, only: [:create, :destroy]
   before_action :set_current_page, only: [:index, :search]
@@ -45,10 +43,6 @@ class Api::V1::Accounts::Companies::ContactsController < Api::V1::Accounts::Ente
     @current_page = params[:page] || 1
   end
 
-  def fetch_company
-    @company = Current.account.companies.find(params[:company_id])
-  end
-
   def fetch_contact
     @contact = @company.contacts.find(params[:id])
   end
@@ -70,18 +64,4 @@ class Api::V1::Accounts::Companies::ContactsController < Api::V1::Accounts::Ente
   def membership_service
     @membership_service ||= Companies::ContactMembershipService.new(company: @company)
   end
-
-  def ensure_companies_enabled!
-    return if Current.account.feature_enabled?('companies')
-
-    render json: { error: 'Companies are not enabled for this account' }, status: :forbidden
-  end
-
-  def authorize_company_read!
-    authorize(@company, :show?)
-  end
-
-  def authorize_company_update!
-    authorize(@company, :update?)
-  end
 end
diff --git a/enterprise/app/controllers/api/v1/accounts/companies/conversations_controller.rb b/enterprise/app/controllers/api/v1/accounts/companies/conversations_controller.rb
new file mode 100644
index 000000000..e6944674e
--- /dev/null
+++ b/enterprise/app/controllers/api/v1/accounts/companies/conversations_controller.rb
@@ -0,0 +1,15 @@
+class Api::V1::Accounts::Companies::ConversationsController < Api::V1::Accounts::Companies::BaseController
+  before_action :authorize_company_read!
+
+  def index
+    conversations = Current.account.conversations.includes(
+      :assignee, :contact, :inbox, :taggings
+    ).where(contact_id: @company.contacts.select(:id))
+
+    @conversations = Conversations::PermissionFilterService.new(
+      conversations,
+      Current.user,
+      Current.account
+    ).perform.order(last_activity_at: :desc).limit(20)
+  end
+end
diff --git a/enterprise/app/controllers/api/v1/accounts/companies/notes_controller.rb b/enterprise/app/controllers/api/v1/accounts/companies/notes_controller.rb
new file mode 100644
index 000000000..85c3df034
--- /dev/null
+++ b/enterprise/app/controllers/api/v1/accounts/companies/notes_controller.rb
@@ -0,0 +1,11 @@
+class Api::V1::Accounts::Companies::NotesController < Api::V1::Accounts::Companies::BaseController
+  before_action :authorize_company_read!
+
+  def index
+    @notes = Current.account.notes
+                    .where(contact_id: @company.contacts.select(:id))
+                    .latest
+                    .includes(:contact, :user)
+                    .limit(20)
+  end
+end
diff --git a/enterprise/app/views/api/v1/accounts/companies/conversations/index.json.jbuilder b/enterprise/app/views/api/v1/accounts/companies/conversations/index.json.jbuilder
new file mode 100644
index 000000000..c81e4bdd0
--- /dev/null
+++ b/enterprise/app/views/api/v1/accounts/companies/conversations/index.json.jbuilder
@@ -0,0 +1,5 @@
+json.payload do
+  json.array! @conversations do |conversation|
+    json.partial! 'api/v1/conversations/partials/conversation', formats: [:json], conversation: conversation
+  end
+end
diff --git a/enterprise/app/views/api/v1/accounts/companies/notes/index.json.jbuilder b/enterprise/app/views/api/v1/accounts/companies/notes/index.json.jbuilder
new file mode 100644
index 000000000..053ebcd97
--- /dev/null
+++ b/enterprise/app/views/api/v1/accounts/companies/notes/index.json.jbuilder
@@ -0,0 +1,8 @@
+json.payload do
+  json.array! @notes do |note|
+    json.partial! 'api/v1/models/note', formats: [:json], resource: note
+    json.contact do
+      json.partial! 'api/v1/models/contact', formats: [:json], resource: note.contact
+    end
+  end
+end

From 85ddc688341747d51b2c75d32a51e8208b666310 Mon Sep 17 00:00:00 2001
From: Sivin Varghese <64252451+iamsivin@users.noreply.github.com>
Date: Tue, 12 May 2026 07:13:48 +0530
Subject: [PATCH 14/20] fix: prevent bulk action checkbox reset in team view
 (#14432)

---
 .../conversationBulkActions/BulkTeamActions.vue           | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)

diff --git a/app/javascript/dashboard/components/widgets/conversation/conversationBulkActions/BulkTeamActions.vue b/app/javascript/dashboard/components/widgets/conversation/conversationBulkActions/BulkTeamActions.vue
index 06a58d961..3e7bc49ec 100644
--- a/app/javascript/dashboard/components/widgets/conversation/conversationBulkActions/BulkTeamActions.vue
+++ b/app/javascript/dashboard/components/widgets/conversation/conversationBulkActions/BulkTeamActions.vue
@@ -1,9 +1,8 @@
 <script setup>
-import { useTemplateRef, computed, ref, onMounted } from 'vue';
+import { useTemplateRef, computed, ref } from 'vue';
 import { useI18n, I18nT } from 'vue-i18n';
 import { useToggle } from '@vueuse/core';
 import { vOnClickOutside } from '@vueuse/components';
-import { useStore } from 'vuex';
 import { useMapGetter } from 'dashboard/composables/store';
 
 import Button from 'dashboard/components-next/button/Button.vue';
@@ -19,7 +18,6 @@ const props = defineProps({
 const emit = defineEmits(['select']);
 
 const { t } = useI18n();
-const store = useStore();
 const containerRef = useTemplateRef('containerRef');
 const [showDropdown, toggleDropdown] = useToggle(false);
 const selectedTeam = ref(null);
@@ -77,10 +75,6 @@ const handleDismiss = () => {
   selectedTeam.value = null;
   toggleDropdown(false);
 };
-
-onMounted(() => {
-  store.dispatch('teams/get');
-});
 </script>
 
 <template>

From 79a7423f9fcb89f2ca36045464c6f991687204bb Mon Sep 17 00:00:00 2001
From: "dependabot[bot]" <49699333+dependabot[bot]@users.noreply.github.com>
Date: Mon, 11 May 2026 18:49:31 -0700
Subject: [PATCH 15/20] chore(deps): bump nokogiri from 1.19.1 to 1.19.3
 (#14410)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Bumps [nokogiri](https://github.com/sparklemotion/nokogiri) from 1.19.1
to 1.19.3.
<details>
<summary>Release notes</summary>
<p><em>Sourced from <a
href="https://github.com/sparklemotion/nokogiri/releases">nokogiri's
releases</a>.</em></p>
<blockquote>
<h2>v1.19.3 / 2026-04-27</h2>
<h3>Fixed / Security</h3>
<ul>
<li>Address exponential regex backtracking in CSS selector tokenizer.
See <a
href="https://github.com/sparklemotion/nokogiri/security/advisories/GHSA-c4rq-3m3g-8wgx">GHSA-c4rq-3m3g-8wgx</a>
for more information.</li>
<li>[CRuby] Address memory leak in
<code>XSLT::Stylesheet#transform</code>. See <a
href="https://github.com/sparklemotion/nokogiri/security/advisories/GHSA-v2fc-qm4h-8hqv">GHSA-v2fc-qm4h-8hqv</a>
for more information.</li>
</ul>
<!-- raw HTML omitted -->

<pre><code>46b89e5d7b9e844c2ee360794240c6ea2a4e6fa0c5892a4ed487db621224b639
nokogiri-1.19.3-aarch64-linux-gnu.gem
8392dfdcd21be7a94dbbe9ccc138dea01b97b24cb2dc02a114ca98bfb1d9a0b7
nokogiri-1.19.3-aarch64-linux-musl.gem
3919d5ffc334ad778a4a9eb88fda7dcb8b1fb58c8a52ac640c6dcd2f038e774f
nokogiri-1.19.3-arm-linux-gnu.gem
9ce1cb6346bb9c67b1550eb537aa183ead91e4b6eadb2f36ade02d8dd2a79fb6
nokogiri-1.19.3-arm-linux-musl.gem
71b9bd424b1b7abc18b05052a1a3cfd3627abdca62be280854cc411791357e42
nokogiri-1.19.3-arm64-darwin.gem
40ea6ebf5cf2005dae1dee26dd557d3afb41fb6de6c9764aca8cf06fdb841db1
nokogiri-1.19.3-java.gem
8bb7132cad356c879a1286eaabcb5e68326cb2490317984280fbc62f456d506a
nokogiri-1.19.3-x64-mingw-ucrt.gem
77f3fba57d46c53ab31e62fc6c28f705109d1bf6264356c76f132b2be5728d4d
nokogiri-1.19.3-x86_64-darwin.gem
2f5078620fe12e83669b5b17311b32532a8153d02eee7ad06948b926d6080976
nokogiri-1.19.3-x86_64-linux-gnu.gem
248c906d2166eca5efb56d52fdee5f9a1f51d69a72e2b64fdac647b4ce39ea3f
nokogiri-1.19.3-x86_64-linux-musl.gem
78312cbac32a40c812780d9678221b79d51288eec00054c1a8d15f7ce05960e8
nokogiri-1.19.3.gem
</code></pre>
<h2>v1.19.2 / 2026-03-19</h2>
<h3>Dependencies</h3>
<ul>
<li>[JRuby] Saxon-HE is updated to 12.7, from 9.6.0-4. Saxon-HE is a
transitive dependency of nu.validator:jing, and this update addresses
CVEs in Saxon-HE's own transitive dependencies JDOM and dom4j. We don't
think this warrants a security release, however we're cutting a patch
release to help users whose security scanners are flagging this. <a
href="https://redirect.github.com/sparklemotion/nokogiri/issues/3611">#3611</a>
<a
href="https://github.com/flavorjones"><code>@​flavorjones</code></a></li>
</ul>
<h3>SHA256 Checksums</h3>

<pre><code>c34d5c8208025587554608e98fd88ab125b29c80f9352b821964e9a5d5cfbd19
nokogiri-1.19.2-aarch64-linux-gnu.gem
7f6b4b0202d507326841a4f790294bf75098aef50c7173443812e3ac5cb06515
nokogiri-1.19.2-aarch64-linux-musl.gem
b7fa1139016f3dc850bda1260988f0d749934a939d04ef2da13bec060d7d5081
nokogiri-1.19.2-arm-linux-gnu.gem
61114d44f6742ff72194a1b3020967201e2eb982814778d130f6471c11f9828c
nokogiri-1.19.2-arm-linux-musl.gem
58d8ea2e31a967b843b70487a44c14c8ba1866daa1b9da9be9dbdf1b43dee205
nokogiri-1.19.2-arm64-darwin.gem
e9d67034bc80ca71043040beea8a91be5dc99b662daa38a2bfb361b7a2cc8717
nokogiri-1.19.2-java.gem
8ccf25eea3363a2c7b3f2e173a3400582c633cfead27f805df9a9c56d4852d1a
nokogiri-1.19.2-x64-mingw-ucrt.gem
7d9af11fda72dfaa2961d8c4d5380ca0b51bc389dc5f8d4b859b9644f195e7a4
nokogiri-1.19.2-x86_64-darwin.gem
fa8feca882b73e871a9845f3817a72e9734c8e974bdc4fbad6e4bc6e8076b94f
nokogiri-1.19.2-x86_64-linux-gnu.gem
93128448e61a9383a30baef041bf1f5817e22f297a1d400521e90294445069a8
nokogiri-1.19.2-x86_64-linux-musl.gem
38fdd8b59db3d5ea9e7dfb14702e882b9bf819198d5bf976f17ebce12c481756
nokogiri-1.19.2.gem
</code></pre>
<p><strong>Full Changelog</strong>: <a
href="https://github.com/sparklemotion/nokogiri/compare/v1.19.1...v1.19.2">https://github.com/sparklemotion/nokogiri/compare/v1.19.1...v1.19.2</a></p>
</blockquote>
</details>
<details>
<summary>Changelog</summary>
<p><em>Sourced from <a
href="https://github.com/sparklemotion/nokogiri/blob/main/CHANGELOG.md">nokogiri's
changelog</a>.</em></p>
<blockquote>
<h2>v1.19.3 / 2026-04-27</h2>
<h3>Fixed / Security</h3>
<ul>
<li>Address exponential regex backtracking in CSS selector tokenizer.
See <a
href="https://github.com/sparklemotion/nokogiri/security/advisories/GHSA-c4rq-3m3g-8wgx">GHSA-c4rq-3m3g-8wgx</a>
for more information.</li>
<li>[CRuby] Address memory leak in
<code>XSLT::Stylesheet#transform</code>. See <a
href="https://github.com/sparklemotion/nokogiri/security/advisories/GHSA-v2fc-qm4h-8hqv">GHSA-v2fc-qm4h-8hqv</a>
for more information.</li>
</ul>
<h2>v1.19.2 / 2026-03-19</h2>
<h3>Dependencies</h3>
<ul>
<li>[JRuby] Saxon-HE is updated to 12.7, from 9.6.0-4. Saxon-HE is a
transitive dependency of nu.validator:jing, and this update addresses
CVEs in Saxon-HE's own transitive dependencies JDOM and dom4j. We don't
think this warrants a security release, however we're cutting a patch
release to help users whose security scanners are flagging this. <a
href="https://redirect.github.com/sparklemotion/nokogiri/issues/3611">#3611</a>
<a
href="https://github.com/flavorjones"><code>@​flavorjones</code></a></li>
</ul>
</blockquote>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a
href="https://github.com/sparklemotion/nokogiri/commit/c139a3da0fe0cae7499a0bafa20f2875877c585b"><code>c139a3d</code></a>
version bump to v1.19.3</li>
<li><a
href="https://github.com/sparklemotion/nokogiri/commit/7501a63b9f4246d12516e35b91fed8be34f854c0"><code>7501a63</code></a>
fix: backtracking in CSS tokenizer rules (v1.19.x backport) (<a
href="https://redirect.github.com/sparklemotion/nokogiri/issues/3627">#3627</a>)</li>
<li><a
href="https://github.com/sparklemotion/nokogiri/commit/03e7968a730a6544ab56a8d6c3e82dd630ad4339"><code>03e7968</code></a>
test: skip CSS tokenizer benchmarks on JRuby</li>
<li><a
href="https://github.com/sparklemotion/nokogiri/commit/b984b7e47f622d1aa97d54c16d5cd596c3eb9538"><code>b984b7e</code></a>
fix: ReDoS in CSS tokenizer ident rule</li>
<li><a
href="https://github.com/sparklemotion/nokogiri/commit/00926231e28d5a20e5b4873efba36099aea0d5c6"><code>0092623</code></a>
fix: ReDoS in CSS tokenizer STRING rule</li>
<li><a
href="https://github.com/sparklemotion/nokogiri/commit/ee17d33aff3adb30c14e71d3d4c8163465acaccf"><code>ee17d33</code></a>
fix: memory leak in XSLT transform (backport to v1.19.x) (<a
href="https://redirect.github.com/sparklemotion/nokogiri/issues/3624">#3624</a>)</li>
<li><a
href="https://github.com/sparklemotion/nokogiri/commit/ce188a395192e3757d8701949afb643dc025084c"><code>ce188a3</code></a>
doc: update CHANGELOG</li>
<li><a
href="https://github.com/sparklemotion/nokogiri/commit/caeaac41f874f0944f9397c78bf6c1bfac2cb472"><code>caeaac4</code></a>
fix: memory leak in XSLT transform</li>
<li><a
href="https://github.com/sparklemotion/nokogiri/commit/25220bf268c9808e28415563ed7f8ea8d5c332bf"><code>25220bf</code></a>
dep(test): test against libxml-ruby v6 (<a
href="https://redirect.github.com/sparklemotion/nokogiri/issues/3618">#3618</a>)</li>
<li><a
href="https://github.com/sparklemotion/nokogiri/commit/0caeb21a5c5e9ff45bbede88fb53655f6753bb0e"><code>0caeb21</code></a>
doc: add security warnings for untrusted XSLT stylesheets</li>
<li>Additional commits viewable in <a
href="https://github.com/sparklemotion/nokogiri/compare/v1.19.1...v1.19.3">compare
view</a></li>
</ul>
</details>
<br />


[![Dependabot compatibility
score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name=nokogiri&package-manager=bundler&previous-version=1.19.1&new-version=1.19.3)](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores)

Dependabot will resolve any conflicts with this PR as long as you don't
alter it yourself. You can also trigger a rebase manually by commenting
`@dependabot rebase`.

[//]: # (dependabot-automerge-start)
[//]: # (dependabot-automerge-end)

---

<details>
<summary>Dependabot commands and options</summary>
<br />

You can trigger Dependabot actions by commenting on this PR:
- `@dependabot rebase` will rebase this PR
- `@dependabot recreate` will recreate this PR, overwriting any edits
that have been made to it
- `@dependabot show <dependency name> ignore conditions` will show all
of the ignore conditions of the specified dependency
- `@dependabot ignore this major version` will close this PR and stop
Dependabot creating any more for this major version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this minor version` will close this PR and stop
Dependabot creating any more for this minor version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this dependency` will close this PR and stop
Dependabot creating any more for this dependency (unless you reopen the
PR or upgrade to it yourself)
You can disable automated security fix PRs for this repo from the
[Security Alerts
page](https://github.com/chatwoot/chatwoot/network/alerts).

</details>

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
---
 Gemfile.lock | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/Gemfile.lock b/Gemfile.lock
index 15d9b2614..7b94e6fe1 100644
--- a/Gemfile.lock
+++ b/Gemfile.lock
@@ -590,14 +590,14 @@ GEM
     newrelic_rpm (9.6.0)
       base64
     nio4r (2.7.3)
-    nokogiri (1.19.1)
+    nokogiri (1.19.3)
       mini_portile2 (~> 2.8.2)
       racc (~> 1.4)
-    nokogiri (1.19.1-arm64-darwin)
+    nokogiri (1.19.3-arm64-darwin)
       racc (~> 1.4)
-    nokogiri (1.19.1-x86_64-darwin)
+    nokogiri (1.19.3-x86_64-darwin)
       racc (~> 1.4)
-    nokogiri (1.19.1-x86_64-linux-gnu)
+    nokogiri (1.19.3-x86_64-linux-gnu)
       racc (~> 1.4)
     oauth (1.1.0)
       oauth-tty (~> 1.0, >= 1.0.1)

From de696a55cbbc0af3ad55e2cffe856ac883d139b6 Mon Sep 17 00:00:00 2001
From: Tanmay Deep Sharma <32020192+tds-1@users.noreply.github.com>
Date: Tue, 12 May 2026 11:23:57 +0530
Subject: [PATCH 16/20] feat(voice): add WhatsApp inbound call webhook pipeline
 [3] (#14315)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Adds the server-side flow that turns Meta WhatsApp Cloud Calling
webhooks into Chatwoot Calls, conversations, voice_call message bubbles,
and ActionCable broadcasts. Stacked on top of #14312 (PR-2 — provider
methods); intentionally does not include the HTTP controller, routes, or
frontend (those land in PR-4 and PR-9).

## Closes
- Part of the WhatsApp Cloud Calling rollout. Linear: TBD

## What changed

**Webhook routing**
- `app/jobs/webhooks/whatsapp_events_job.rb` — append
`prepend_mod_with('Webhooks::WhatsappEventsJob')` so EE can extend it
without forking.
- `enterprise/app/jobs/enterprise/webhooks/whatsapp_events_job.rb` (new)
— overlay that prepends `handle_message_events` to intercept `field:
'calls'` payloads (route to `Whatsapp::IncomingCallService`) and
`interactive.call_permission_reply` messages (route to
`Whatsapp::CallPermissionReplyService`); falls through with `super` for
regular messages.

**Services**
- `enterprise/app/services/whatsapp/incoming_call_service.rb` (new) —
gated on `provider_config['calling_enabled']`; processes `connect`
(creates inbound call via `Voice::InboundCallBuilder` or transitions an
existing outbound call to `in_progress`) and `terminate` events; updates
conversation `additional_attributes` and broadcasts
`voice_call.incoming`/`voice_call.outbound_connected`/`voice_call.ended`.
- `enterprise/app/services/whatsapp/call_permission_reply_service.rb`
(new) — handles WhatsApp interactive `call_permission_reply` replies;
clears the conversation's `call_permission_requested_at` flag and
broadcasts `voice_call.permission_granted` so the agent UI can re-enable
the call button.

**Builder/model adjustments**
- `enterprise/app/services/voice/inbound_call_builder.rb` —
provider-agnostic; accepts `provider:` and `extra_meta:` kwargs, drops
`account:` (now derived from `inbox.account` to keep the param count
under rubocop's ceiling without disabling cops), uses digits-only
`source_id` for WhatsApp ContactInbox (validation requires
`^\d{1,15}\z`), skips Twilio-only `conference_sid` for non-Twilio
providers.
- `enterprise/app/services/voice/call_message_builder.rb` — adds
`create!`/`update_status!` API and `CALL_TO_VOICE_STATUS` map; uses
direct `Message.create!` (bypasses `Messages::MessageBuilder`'s
incoming-on-non-Api-inbox guard, which would otherwise reject the system
bubble); content is `'WhatsApp Call'` for WhatsApp and `'Voice Call'`
for Twilio. Backwards-compatible `perform!` retained for the existing
Twilio call sites.
- `enterprise/app/models/call.rb` — adds `default_ice_servers` (driven
by `VOICE_CALL_STUN_URLS` env), `direction_label` alias for the
`inbound`/`outbound` strings the FE expects, and
`ringing?`/`in_progress?`/`terminal?` predicates used throughout the
pipeline.

**Outgoing-channel guard**
- `app/services/base/send_on_channel_service.rb` — extends
`invalid_message?` to skip messages with `content_type == 'voice_call'`.
Without this, agent-initiated outbound calls (PR-4) would deliver
\"WhatsApp Call\" as a text message to the contact every time.

**Twilio call-site update**
- `enterprise/app/controllers/twilio/voice_controller.rb` — drops the
now-redundant `account: current_account` kwarg from the
`Voice::InboundCallBuilder.perform!` call.

**Tests**
- New: `spec/enterprise/services/whatsapp/incoming_call_service_spec.rb`
(5 examples — calling-disabled, inbound connect, outbound connect,
terminate completed, terminate no-answer, unknown event).
- New:
`spec/enterprise/services/whatsapp/call_permission_reply_service_spec.rb`
(3 examples — accept, reject, calling-disabled).
- Updated: `spec/enterprise/services/voice/inbound_call_builder_spec.rb`
and `spec/enterprise/controllers/twilio/voice_controller_spec.rb` to
drop the `account:` kwarg from call expectations.

## How to test

In `rails console` against an account with a WhatsApp inbox where
`provider_config['calling_enabled']` is true:

```ruby
inbox = Inbox.find(<id>)
params = { calls: [{ id: 'wacid_test', from: '15550001111', event: 'connect',
                     session: { sdp: 'v=0...', sdp_type: 'offer' } }] }
Whatsapp::IncomingCallService.new(inbox: inbox, params: params).perform
# => Conversation + Call (status: 'ringing', provider: 'whatsapp') + voice_call message bubble
# => ActionCable broadcasts `voice_call.incoming` to the assignee or account-wide

# Then terminate it:
Whatsapp::IncomingCallService.new(inbox: inbox,
  params: { calls: [{ id: 'wacid_test', event: 'terminate', duration: 0, terminate_reason: 'no_answer' }] }
).perform
# => Call status flips to 'no_answer', message bubble updates, `voice_call.ended` broadcast fires
```

End-to-end browser flow (Meta → cable → UI) requires the controller from
PR-4 and the frontend from PR-9.

---------

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---
 .../messages/messenger/message_builder.rb     |   4 +
 app/jobs/webhooks/whatsapp_events_job.rb      |   2 +
 app/models/attachment.rb                      |  12 +
 app/models/channel/whatsapp.rb                |  10 +
 app/services/base/send_on_channel_service.rb  |   3 +-
 app/services/whatsapp/facebook_api_client.rb  |   2 +-
 app/views/api/v1/models/_inbox.json.jbuilder  |   3 +
 config/initializers/active_storage.rb         |  13 +
 config/locales/en.yml                         |  13 +
 config/routes.rb                              |  15 ++
 .../enterprise/messages/message_builder.rb    |   7 +-
 .../v1/accounts/contacts/calls_controller.rb  |  19 +-
 .../v1/accounts/whatsapp_calls_controller.rb  | 187 ++++++++++++++
 .../controllers/twilio/voice_controller.rb    |   1 -
 .../webhooks/whatsapp_events_job.rb           |  59 +++++
 enterprise/app/models/call.rb                 |  29 ++-
 .../models/enterprise/concerns/attachment.rb  |  19 ++
 .../app/models/enterprise/conversation.rb     |  15 ++
 .../providers/whatsapp_cloud_service.rb       |  19 +-
 .../services/voice/call_message_builder.rb    |  30 ++-
 .../services/voice/inbound_call_builder.rb    |  61 +++--
 .../services/voice/outbound_call_builder.rb   |   9 +-
 .../whatsapp/call_permission_reply_service.rb |  73 ++++++
 .../app/services/whatsapp/call_service.rb     | 110 ++++++++
 .../whatsapp/incoming_call_service.rb         | 186 ++++++++++++++
 .../whatsapp_calls/accept.json.jbuilder       |   1 +
 .../whatsapp_calls/initiate.json.jbuilder     |   5 +
 .../whatsapp_calls/reject.json.jbuilder       |   2 +
 .../whatsapp_calls/show.json.jbuilder         |   1 +
 .../whatsapp_calls/terminate.json.jbuilder    |   2 +
 .../upload_recording.json.jbuilder            |   2 +
 .../v1/models/_whatsapp_call.json.jbuilder    |  24 ++
 .../twilio/voice_controller_spec.rb           |   1 -
 .../providers/whatsapp_cloud_service_spec.rb  |   8 +-
 .../voice/inbound_call_builder_spec.rb        |  40 ++-
 .../call_permission_reply_service_spec.rb     |  98 +++++++
 .../whatsapp/incoming_call_service_spec.rb    | 242 ++++++++++++++++++
 .../telegram/send_attachments_service_spec.rb |   4 +-
 .../whatsapp/facebook_api_client_spec.rb      |   4 +-
 39 files changed, 1288 insertions(+), 47 deletions(-)
 create mode 100644 config/initializers/active_storage.rb
 create mode 100644 enterprise/app/controllers/api/v1/accounts/whatsapp_calls_controller.rb
 create mode 100644 enterprise/app/jobs/enterprise/webhooks/whatsapp_events_job.rb
 create mode 100644 enterprise/app/services/whatsapp/call_permission_reply_service.rb
 create mode 100644 enterprise/app/services/whatsapp/call_service.rb
 create mode 100644 enterprise/app/services/whatsapp/incoming_call_service.rb
 create mode 100644 enterprise/app/views/api/v1/accounts/whatsapp_calls/accept.json.jbuilder
 create mode 100644 enterprise/app/views/api/v1/accounts/whatsapp_calls/initiate.json.jbuilder
 create mode 100644 enterprise/app/views/api/v1/accounts/whatsapp_calls/reject.json.jbuilder
 create mode 100644 enterprise/app/views/api/v1/accounts/whatsapp_calls/show.json.jbuilder
 create mode 100644 enterprise/app/views/api/v1/accounts/whatsapp_calls/terminate.json.jbuilder
 create mode 100644 enterprise/app/views/api/v1/accounts/whatsapp_calls/upload_recording.json.jbuilder
 create mode 100644 enterprise/app/views/api/v1/models/_whatsapp_call.json.jbuilder
 create mode 100644 spec/enterprise/services/whatsapp/call_permission_reply_service_spec.rb
 create mode 100644 spec/enterprise/services/whatsapp/incoming_call_service_spec.rb

diff --git a/app/builders/messages/messenger/message_builder.rb b/app/builders/messages/messenger/message_builder.rb
index be1f98b43..ecd6f06ea 100644
--- a/app/builders/messages/messenger/message_builder.rb
+++ b/app/builders/messages/messenger/message_builder.rb
@@ -28,6 +28,10 @@ class Messages::Messenger::MessageBuilder
       filename: attachment_file.original_filename,
       content_type: attachment_file.content_type
     )
+    # The Attachment row is saved before the blob is attached, so the
+    # after_create_commit broadcast bails on `file.attached?`. Re-fire here
+    # for audio so the bubble updates without waiting on transcription.
+    attachment.message&.reload&.send_update_event if attachment.file_type.to_sym == :audio
   end
 
   def attachment_params(attachment)
diff --git a/app/jobs/webhooks/whatsapp_events_job.rb b/app/jobs/webhooks/whatsapp_events_job.rb
index b0bede8b0..49b7265b8 100644
--- a/app/jobs/webhooks/whatsapp_events_job.rb
+++ b/app/jobs/webhooks/whatsapp_events_job.rb
@@ -128,3 +128,5 @@ class Webhooks::WhatsappEventsJob < MutexApplicationJob
     return channel if channel && channel.provider_config['phone_number_id'] == phone_number_id
   end
 end
+
+Webhooks::WhatsappEventsJob.prepend_mod_with('Webhooks::WhatsappEventsJob')
diff --git a/app/models/attachment.rb b/app/models/attachment.rb
index 769e134be..8c87ff433 100644
--- a/app/models/attachment.rb
+++ b/app/models/attachment.rb
@@ -104,11 +104,23 @@ class Attachment < ApplicationRecord
     audio_file_data = base_data.merge(file_metadata)
     audio_file_data.merge(
       {
+        # ActiveStorage's redirect endpoint defaults to Content-Disposition: attachment,
+        # which makes <audio> elements download instead of play. Force inline so the
+        # call-recording chip (and any other audio bubble) can stream directly.
+        data_url: inline_audio_url,
         transcribed_text: meta&.[]('transcribed_text') || ''
       }
     )
   end
 
+  def inline_audio_url
+    return '' unless file.attached?
+
+    # Proxy endpoint streams through Rails and honours `disposition: 'inline'`,
+    # unlike the redirect endpoint which always sends Content-Disposition: attachment.
+    Rails.application.routes.url_helpers.rails_storage_proxy_url(file, disposition: 'inline')
+  end
+
   def file_metadata
     metadata = {
       extension: extension,
diff --git a/app/models/channel/whatsapp.rb b/app/models/channel/whatsapp.rb
index 5905c54f7..e1d4b226a 100644
--- a/app/models/channel/whatsapp.rb
+++ b/app/models/channel/whatsapp.rb
@@ -40,6 +40,16 @@ class Channel::Whatsapp < ApplicationRecord
     'Whatsapp'
   end
 
+  # Mirrors Channel::TwilioSms#voice_enabled? so the call subsystem can duck-type across providers.
+  # Meta's Calling API is only available via the embedded-signup whatsapp_cloud flow —
+  # 360dialog (default provider) and manual whatsapp_cloud setups can't reach the call APIs.
+  def voice_enabled?
+    provider == 'whatsapp_cloud' &&
+      provider_config['source'] == 'embedded_signup' &&
+      provider_config['calling_enabled'].present? &&
+      account.feature_enabled?('channel_voice')
+  end
+
   def provider_service
     if provider == 'whatsapp_cloud'
       Whatsapp::Providers::WhatsappCloudService.new(whatsapp_channel: self)
diff --git a/app/services/base/send_on_channel_service.rb b/app/services/base/send_on_channel_service.rb
index 298f60962..44380a179 100644
--- a/app/services/base/send_on_channel_service.rb
+++ b/app/services/base/send_on_channel_service.rb
@@ -46,7 +46,8 @@ class Base::SendOnChannelService
   def invalid_message?
     # private notes aren't send to the channels
     # we should also avoid the case of message loops, when outgoing messages are created from channel
-    message.private? || outgoing_message_originated_from_channel?
+    # voice_call bubbles are call status indicators, not deliverable messages
+    message.private? || outgoing_message_originated_from_channel? || message.content_type == 'voice_call'
   end
 
   def validate_target_channel
diff --git a/app/services/whatsapp/facebook_api_client.rb b/app/services/whatsapp/facebook_api_client.rb
index fa09a4b44..94e46dabd 100644
--- a/app/services/whatsapp/facebook_api_client.rb
+++ b/app/services/whatsapp/facebook_api_client.rb
@@ -86,7 +86,7 @@ class Whatsapp::FacebookApiClient
       body: {
         override_callback_uri: callback_url,
         verify_token: verify_token,
-        subscribed_fields: %w[messages smb_message_echoes]
+        subscribed_fields: %w[messages smb_message_echoes calls]
       }.to_json
     )
 
diff --git a/app/views/api/v1/models/_inbox.json.jbuilder b/app/views/api/v1/models/_inbox.json.jbuilder
index 3da71d07d..0ae0745cd 100644
--- a/app/views/api/v1/models/_inbox.json.jbuilder
+++ b/app/views/api/v1/models/_inbox.json.jbuilder
@@ -143,3 +143,6 @@ if resource.twilio? && resource.channel.respond_to?(:voice_enabled?)
     json.voice_status_webhook_url resource.channel.try(:voice_status_webhook_url)
   end
 end
+
+## Voice attribute for WhatsApp Cloud (only embedded-signup channels surface true)
+json.voice_enabled resource.channel.voice_enabled? if resource.channel_type == 'Channel::Whatsapp' && resource.channel.respond_to?(:voice_enabled?)
diff --git a/config/initializers/active_storage.rb b/config/initializers/active_storage.rb
new file mode 100644
index 000000000..d927e1472
--- /dev/null
+++ b/config/initializers/active_storage.rb
@@ -0,0 +1,13 @@
+# Allow audio attachments (call recordings, voice notes) to serve inline so the
+# in-app <audio> player can stream them. Without this, ActiveStorage's blob model
+# forces Content-Disposition: attachment for any MIME outside the default allowlist
+# (images + PDF), which makes the browser download instead of play.
+Rails.application.config.active_storage.content_types_allowed_inline += %w[
+  audio/webm
+  audio/ogg
+  audio/mpeg
+  audio/mp4
+  audio/x-m4a
+  audio/wav
+  audio/x-wav
+]
diff --git a/config/locales/en.yml b/config/locales/en.yml
index 8a53e12c5..f7b98a47b 100644
--- a/config/locales/en.yml
+++ b/config/locales/en.yml
@@ -116,6 +116,13 @@ en:
       reauthorization:
         generic: 'Failed to reauthorize WhatsApp. Please try again.'
         not_supported: 'Reauthorization is not supported for this type of WhatsApp channel.'
+      calls:
+        not_enabled: 'Calling is not enabled for this inbox'
+        no_recording: 'No recording file provided'
+        no_message: 'Call has no associated message'
+        sdp_offer_required: 'sdp_offer is required'
+        contact_phone_required: 'Contact phone number is required'
+        permission_request_failed: 'Failed to send call permission request'
     inboxes:
       imap:
         socket_error: Please check the network connection, IMAP address and try again.
@@ -246,6 +253,9 @@ en:
       whatsapp:
         list_button_label: 'Choose an item'
         call_permission_request_body: 'We would like to call you regarding your conversation.'
+      voice_call:
+        twilio: 'Voice Call'
+        whatsapp: 'WhatsApp Call'
       delivery_status:
         error_code: 'Error code: %{error_code}'
     activity:
@@ -290,6 +300,9 @@ en:
         issue_created: 'Linear issue %{issue_id} was created by %{user_name}'
         issue_linked: 'Linear issue %{issue_id} was linked by %{user_name}'
         issue_unlinked: 'Linear issue %{issue_id} was unlinked by %{user_name}'
+      whatsapp_call:
+        permission_requested: 'Sent a call permission request to %{contact_name}.'
+        permission_granted: '%{contact_name} accepted the call permission request.'
       csat:
         not_sent_due_to_messaging_window: 'CSAT survey not sent due to outgoing message restrictions'
       auto_resolve:
diff --git a/config/routes.rb b/config/routes.rb
index ba6a33940..dec54c093 100644
--- a/config/routes.rb
+++ b/config/routes.rb
@@ -228,6 +228,21 @@ Rails.application.routes.draw do
             end
           end
           resources :reporting_events, only: [:index] if ChatwootApp.enterprise?
+
+          if ChatwootApp.enterprise?
+            resources :whatsapp_calls, only: [:show] do
+              member do
+                post :accept
+                post :reject
+                post :terminate
+                post :upload_recording
+              end
+              collection do
+                post :initiate
+              end
+            end
+          end
+
           resources :custom_attribute_definitions, only: [:index, :show, :create, :update, :destroy]
           resources :custom_filters, only: [:index, :show, :create, :update, :destroy]
           resources :inboxes, only: [:index, :show, :create, :update, :destroy] do
diff --git a/enterprise/app/builders/enterprise/messages/message_builder.rb b/enterprise/app/builders/enterprise/messages/message_builder.rb
index 06bd0357c..2240b4b43 100644
--- a/enterprise/app/builders/enterprise/messages/message_builder.rb
+++ b/enterprise/app/builders/enterprise/messages/message_builder.rb
@@ -2,13 +2,12 @@ module Enterprise::Messages::MessageBuilder
   private
 
   def message_type
-    return @message_type if @message_type == 'incoming' && twilio_voice_inbox? && @params[:content_type] == 'voice_call'
+    return @message_type if @message_type == 'incoming' && voice_call_inbox? && @params[:content_type] == 'voice_call'
 
     super
   end
 
-  def twilio_voice_inbox?
-    inbox = @conversation.inbox
-    inbox.channel_type == 'Channel::TwilioSms' && inbox.channel.voice_enabled?
+  def voice_call_inbox?
+    @conversation.inbox.channel.try(:voice_enabled?)
   end
 end
diff --git a/enterprise/app/controllers/api/v1/accounts/contacts/calls_controller.rb b/enterprise/app/controllers/api/v1/accounts/contacts/calls_controller.rb
index 8217e1a37..740a2fda4 100644
--- a/enterprise/app/controllers/api/v1/accounts/contacts/calls_controller.rb
+++ b/enterprise/app/controllers/api/v1/accounts/contacts/calls_controller.rb
@@ -10,7 +10,8 @@ class Api::V1::Accounts::Contacts::CallsController < Api::V1::Accounts::BaseCont
       account: Current.account,
       inbox: voice_inbox,
       user: Current.user,
-      contact: contact
+      contact: contact,
+      conversation: existing_conversation
     )
 
     render json: {
@@ -38,4 +39,20 @@ class Api::V1::Accounts::Contacts::CallsController < Api::V1::Accounts::BaseCont
       inbox
     end
   end
+
+  # Reuse the open conversation when the caller is already inside it; ignore
+  # the hint if it doesn't belong to the picked voice inbox or dialed contact.
+  # Only reuse open conversations — Message#reopen_conversation skips outgoing
+  # messages, so dropping a voice_call bubble into a resolved/snoozed/pending
+  # thread leaves it stuck in that state.
+  def existing_conversation
+    return nil if params[:conversation_id].blank?
+
+    conversation = Current.account.conversations.find_by(display_id: params[:conversation_id])
+    return nil unless conversation
+    return nil unless conversation.inbox_id == voice_inbox.id && conversation.contact_id == contact.id
+    return nil unless conversation.open?
+
+    conversation
+  end
 end
diff --git a/enterprise/app/controllers/api/v1/accounts/whatsapp_calls_controller.rb b/enterprise/app/controllers/api/v1/accounts/whatsapp_calls_controller.rb
new file mode 100644
index 000000000..b54940586
--- /dev/null
+++ b/enterprise/app/controllers/api/v1/accounts/whatsapp_calls_controller.rb
@@ -0,0 +1,187 @@
+class Api::V1::Accounts::WhatsappCallsController < Api::V1::Accounts::BaseController
+  PERMISSION_REQUEST_THROTTLE = 5.minutes
+
+  before_action :set_call, only: %i[show accept reject terminate upload_recording]
+  before_action :set_conversation, only: :initiate
+  before_action :ensure_calling_enabled, only: :initiate
+  before_action :ensure_sdp_offer, only: :initiate
+  before_action :ensure_contact_phone, only: :initiate
+  before_action :ensure_recording_present, only: :upload_recording
+  before_action :ensure_call_message, only: :upload_recording
+
+  rescue_from Voice::CallErrors::NotRinging,
+              Voice::CallErrors::AlreadyAccepted,
+              Voice::CallErrors::CallFailed,
+              with: :render_call_error
+  rescue_from Voice::CallErrors::NoCallPermission, with: :render_permission_request
+
+  def show; end
+
+  def accept
+    call_service.accept
+  end
+
+  def reject
+    call_service.reject
+  end
+
+  def terminate
+    call_service.terminate
+  end
+
+  def upload_recording
+    @upload_status = @call.message.with_lock { attach_recording_idempotently }
+  end
+
+  def initiate
+    @call = create_outbound_call
+    @message = Voice::CallMessageBuilder.new(@call).perform!
+    @call.update!(message_id: @message.id)
+  end
+
+  private
+
+  def call_service
+    @call_service ||= Whatsapp::CallService.new(call: @call, agent: Current.user, sdp_answer: params[:sdp_answer])
+  end
+
+  def provider_service
+    @provider_service ||= @conversation.inbox.channel.provider_service
+  end
+
+  def set_call
+    @call = Current.account.calls.whatsapp.find(params[:id])
+    authorize @call.conversation, :show?
+  end
+
+  def set_conversation
+    @conversation = Current.account.conversations.find_by!(display_id: params[:conversation_id])
+    authorize @conversation, :show?
+  end
+
+  def ensure_calling_enabled
+    channel = @conversation.inbox.channel
+    return if channel.is_a?(Channel::Whatsapp) && channel.voice_enabled?
+
+    render_could_not_create_error(I18n.t('errors.whatsapp.calls.not_enabled'))
+  end
+
+  def ensure_sdp_offer
+    return if params[:sdp_offer].present?
+
+    render_could_not_create_error(I18n.t('errors.whatsapp.calls.sdp_offer_required'))
+  end
+
+  def ensure_contact_phone
+    return if @conversation.contact&.phone_number.present?
+
+    render_could_not_create_error(I18n.t('errors.whatsapp.calls.contact_phone_required'))
+  end
+
+  def ensure_recording_present
+    return if params[:recording].present?
+
+    render_could_not_create_error(I18n.t('errors.whatsapp.calls.no_recording'))
+  end
+
+  def ensure_call_message
+    return if @call.message.present?
+
+    render_could_not_create_error(I18n.t('errors.whatsapp.calls.no_message'))
+  end
+
+  def attach_recording_idempotently
+    return 'already_uploaded' if @call.message.attachments.exists?(file_type: :audio)
+
+    @call.message.attachments.create!(account_id: @call.account_id, file_type: :audio, file: params[:recording])
+    'uploaded'
+  end
+
+  def create_outbound_call
+    contact_phone = @conversation.contact.phone_number.delete('+')
+    result = provider_service.initiate_call(contact_phone, params[:sdp_offer])
+    provider_call_id = result.dig('calls', 0, 'id') || result['call_id']
+
+    Current.account.calls.create!(
+      provider: :whatsapp, inbox: @conversation.inbox, conversation: @conversation, contact: @conversation.contact,
+      provider_call_id: provider_call_id, direction: :outgoing, status: 'ringing',
+      accepted_by_agent_id: Current.user.id,
+      meta: { 'sdp_offer' => params[:sdp_offer], 'ice_servers' => Call.default_ice_servers }
+    )
+  end
+
+  # Meta error 138006 means the contact hasn't opted in yet; send the opt-in
+  # template (throttled, behind a conversation lock to prevent double-send).
+  def render_permission_request
+    status = nil
+    @conversation.with_lock do
+      if permission_request_throttled?
+        status = 'permission_pending'
+        next
+      end
+
+      sent = send_permission_request_safely
+      if sent
+        record_permission_request_wamid(sent)
+        emit_permission_requested_activity
+        status = 'permission_requested'
+      else
+        status = 'failed'
+      end
+    end
+
+    return render_could_not_create_error(I18n.t('errors.whatsapp.calls.permission_request_failed')) if status == 'failed'
+
+    # 422 (not 200) so any client treating 2xx as "call placed" can't mistake
+    # the permission-template path for a successful dial. The FE composable
+    # detects this status and surfaces the banner instead of throwing.
+    render json: { status: status }, status: :unprocessable_entity
+  end
+
+  def permission_request_throttled?
+    last_requested = @conversation.additional_attributes&.dig('call_permission_requested_at')
+    last_requested.present? && Time.zone.parse(last_requested) > PERMISSION_REQUEST_THROTTLE.ago
+  end
+
+  # Treat transport errors as a falsy return so we render 422 rather than 500.
+  def send_permission_request_safely
+    provider_service.send_call_permission_request(
+      @conversation.contact.phone_number.delete('+'),
+      *permission_request_body_args
+    )
+  rescue StandardError => e
+    Rails.logger.warn "[WHATSAPP CALL] permission_request failed: #{e.class} #{e.message}"
+    nil
+  end
+
+  # Pass the inbox-level override only when present so the provider falls back
+  # to the i18n default for inboxes that haven't customized the prompt.
+  def permission_request_body_args
+    custom_body = @conversation.inbox.channel.provider_config&.dig('call_permission_request_body').presence
+    custom_body ? [custom_body] : []
+  end
+
+  def emit_permission_requested_activity
+    content = I18n.t(
+      'conversations.activity.whatsapp_call.permission_requested',
+      contact_name: @conversation.contact.name
+    )
+    ::Conversations::ActivityMessageJob.perform_later(
+      @conversation,
+      { account_id: @conversation.account_id, inbox_id: @conversation.inbox_id, message_type: :activity, content: content }
+    )
+  end
+
+  # Stash the outbound wamid so the reply webhook can match context.id back here.
+  def record_permission_request_wamid(sent)
+    attrs = (@conversation.additional_attributes || {}).merge(
+      'call_permission_requested_at' => Time.current.iso8601,
+      'call_permission_request_message_id' => sent.dig('messages', 0, 'id')
+    )
+    @conversation.update!(additional_attributes: attrs)
+  end
+
+  def render_call_error(error)
+    render_could_not_create_error(error.message)
+  end
+end
diff --git a/enterprise/app/controllers/twilio/voice_controller.rb b/enterprise/app/controllers/twilio/voice_controller.rb
index 37317d9de..eb29c00bd 100644
--- a/enterprise/app/controllers/twilio/voice_controller.rb
+++ b/enterprise/app/controllers/twilio/voice_controller.rb
@@ -94,7 +94,6 @@ class Twilio::VoiceController < ApplicationController
     case twilio_direction
     when 'inbound'
       Voice::InboundCallBuilder.perform!(
-        account: current_account,
         inbox: inbox,
         from_number: twilio_from,
         call_sid: twilio_call_sid
diff --git a/enterprise/app/jobs/enterprise/webhooks/whatsapp_events_job.rb b/enterprise/app/jobs/enterprise/webhooks/whatsapp_events_job.rb
new file mode 100644
index 000000000..b1e2bc053
--- /dev/null
+++ b/enterprise/app/jobs/enterprise/webhooks/whatsapp_events_job.rb
@@ -0,0 +1,59 @@
+module Enterprise::Webhooks::WhatsappEventsJob
+  def handle_message_events(channel, params)
+    return handle_call_events(channel, params) if call_event?(params)
+    return handle_call_permission_reply(channel, params) if call_permission_reply?(params)
+
+    super
+  end
+
+  private
+
+  # Lock per-call_id inside handle_call_events instead of the parent's per-sender mutex.
+  def contact_sender_id(params)
+    return nil if call_event?(params)
+
+    super
+  end
+
+  def call_event?(params)
+    params.dig(:entry, 0, :changes, 0, :field) == 'calls'
+  end
+
+  def call_permission_reply?(params)
+    params.dig(:entry, 0, :changes, 0, :value, :messages, 0, :interactive, :type) == 'call_permission_reply'
+  end
+
+  # Per-call_id mutex so connect/status/terminate for the same call serialize
+  # across batches. Meta delivers two payload shapes under field=calls:
+  #   - value.calls[]    → event-based (connect, terminate)
+  #   - value.statuses[] → status-based (RINGING, ACCEPTED) — the real pickup
+  #     signal for outbound; without this, only `connect` (tunnel-up) is seen
+  #     and timer/recorder kick off before the contact actually answers.
+  def handle_call_events(channel, params)
+    value = params.dig(:entry, 0, :changes, 0, :value) || {}
+
+    Array(value[:calls]).each do |call_payload|
+      with_call_lock(channel, call_payload[:id]) do
+        Whatsapp::IncomingCallService.new(inbox: channel.inbox, params: { calls: [call_payload] }).perform
+      end
+    end
+
+    Array(value[:statuses]).each do |status_payload|
+      next unless status_payload[:type] == 'call'
+
+      with_call_lock(channel, status_payload[:id]) do
+        Whatsapp::IncomingCallService.new(inbox: channel.inbox, params: { statuses: [status_payload] }).perform
+      end
+    end
+  end
+
+  def with_call_lock(channel, call_id, &)
+    lock_key = format(::Redis::Alfred::WHATSAPP_MESSAGE_MUTEX,
+                      inbox_id: channel.inbox.id, sender_id: "call:#{call_id}")
+    with_lock(lock_key, 30.seconds, &)
+  end
+
+  def handle_call_permission_reply(channel, params)
+    Whatsapp::CallPermissionReplyService.new(inbox: channel.inbox, params: params).perform
+  end
+end
diff --git a/enterprise/app/models/call.rb b/enterprise/app/models/call.rb
index e5c98baa1..f8c0580f8 100644
--- a/enterprise/app/models/call.rb
+++ b/enterprise/app/models/call.rb
@@ -29,13 +29,16 @@
 #  index_calls_on_provider_and_provider_call_id   (provider,provider_call_id) UNIQUE
 #
 class Call < ApplicationRecord
-  # All valid call statuses
   STATUSES = %w[ringing in_progress completed no_answer failed].freeze
-  # Statuses where the call is finished and won't change again
   TERMINAL_STATUSES = %w[completed no_answer failed].freeze
 
   store_accessor :meta, :conference_sid, :twilio_conference_sid, :recording_sid, :parent_call_sid, :initiated_at, :ended_at
 
+  # Frontend voice bubbles/stores expect inbound/outbound string values
+  DISPLAY_DIRECTION = { 'incoming' => 'inbound', 'outgoing' => 'outbound' }.freeze
+
+  DEFAULT_STUN_URL = 'stun:stun.l.google.com:19302'.freeze
+
   enum :provider, { twilio: 0, whatsapp: 1 }
   enum :direction, { incoming: 0, outgoing: 1 }
 
@@ -65,6 +68,28 @@ class Call < ApplicationRecord
     "conf_account_#{account_id}_call_#{id}"
   end
 
+  # Browser ↔ Meta WebRTC needs at least one STUN server to discover its public srflx candidate.
+  def self.default_ice_servers
+    urls = ENV.fetch('VOICE_CALL_STUN_URLS', DEFAULT_STUN_URL).split(',').filter_map { |u| u.strip.presence }
+    [{ urls: urls }]
+  end
+
+  def direction_label
+    DISPLAY_DIRECTION[direction]
+  end
+
+  def ringing?
+    status == 'ringing'
+  end
+
+  def in_progress?
+    status == 'in_progress'
+  end
+
+  def terminal?
+    TERMINAL_STATUSES.include?(status)
+  end
+
   def display_status
     status.to_s.tr('_', '-')
   end
diff --git a/enterprise/app/models/enterprise/concerns/attachment.rb b/enterprise/app/models/enterprise/concerns/attachment.rb
index 155dfcaad..21a13272b 100644
--- a/enterprise/app/models/enterprise/concerns/attachment.rb
+++ b/enterprise/app/models/enterprise/concerns/attachment.rb
@@ -3,6 +3,11 @@ module Enterprise::Concerns::Attachment
 
   included do
     after_create_commit :enqueue_audio_transcription
+    # Broadcast the message update so the FE bubble picks up the new audio
+    # attachment immediately. Without this, the FE has to wait until Whisper
+    # finishes (or fall back to a page refresh) — and if Whisper returns blank,
+    # the bubble never gets the audio at all.
+    after_create_commit :broadcast_message_update_for_audio
   end
 
   private
@@ -10,6 +15,20 @@ module Enterprise::Concerns::Attachment
   def enqueue_audio_transcription
     return unless file_type.to_sym == :audio
 
+    # No file.attached? guard: the social-media ingest path saves the
+    # Attachment before attaching the blob. AudioTranscriptionJob retries
+    # on ActiveStorage::FileNotFoundError to ride out that race.
     Messages::AudioTranscriptionJob.perform_later(id)
   end
+
+  def broadcast_message_update_for_audio
+    return unless file_type.to_sym == :audio
+    return unless message
+    # Without an attached file, the message serializer's audio_metadata path
+    # dereferences `file.metadata[:width]` on nil and raises. The pre-attach
+    # broadcast wouldn't carry useful audio info anyway — skip until upload completes.
+    return unless file.attached?
+
+    message.reload.send_update_event
+  end
 end
diff --git a/enterprise/app/models/enterprise/conversation.rb b/enterprise/app/models/enterprise/conversation.rb
index 09cb2c569..077e0e932 100644
--- a/enterprise/app/models/enterprise/conversation.rb
+++ b/enterprise/app/models/enterprise/conversation.rb
@@ -13,6 +13,12 @@ module Enterprise::Conversation
     super + %w[sla_policy_id]
   end
 
+  # Surface call lifecycle changes to the FE: writes to additional_attributes
+  # call_status/call_direction should rebroadcast conversation_updated.
+  def allowed_keys?
+    super || call_attributes_changed?
+  end
+
   def with_captain_activity_context(reason:, reason_type:)
     previous_reason = captain_activity_reason
     previous_reason_type = captain_activity_reason_type
@@ -30,4 +36,13 @@ module Enterprise::Conversation
   def dispatch_captain_inference_event(event_name)
     dispatcher_dispatch(event_name)
   end
+
+  def call_attributes_changed?
+    return false if previous_changes['additional_attributes'].blank?
+
+    # Compare before/after values for call keys — checking key presence alone
+    # rebroadcasts on any unrelated additional_attributes write once the keys exist.
+    before, after = previous_changes['additional_attributes']
+    %w[call_status call_direction].any? { |key| (before || {})[key] != (after || {})[key] }
+  end
 end
diff --git a/enterprise/app/services/enterprise/whatsapp/providers/whatsapp_cloud_service.rb b/enterprise/app/services/enterprise/whatsapp/providers/whatsapp_cloud_service.rb
index 552c3c389..ec29a5a38 100644
--- a/enterprise/app/services/enterprise/whatsapp/providers/whatsapp_cloud_service.rb
+++ b/enterprise/app/services/enterprise/whatsapp/providers/whatsapp_cloud_service.rb
@@ -1,4 +1,9 @@
 module Enterprise::Whatsapp::Providers::WhatsappCloudService
+  # Calls API + the call_permission_request interactive message both require Graph
+  # API v17+; OSS phone_id_path is locked at v13.0 for legacy /messages compatibility.
+  # Use the configured global version (defaulting to v22.0) for call-flow endpoints.
+  WHATSAPP_CALLING_API_VERSION_FALLBACK = 'v22.0'.freeze
+
   def pre_accept_call(call_id, sdp_answer)
     call_api('pre_accept_call', call_action_body(call_id, 'pre_accept', sdp_answer))
   end
@@ -17,7 +22,7 @@ module Enterprise::Whatsapp::Providers::WhatsappCloudService
 
   def send_call_permission_request(to_phone_number, body_text = I18n.t('conversations.messages.whatsapp.call_permission_request_body'))
     response = HTTParty.post(
-      "#{phone_id_path}/messages", headers: api_headers, body: permission_request_body(to_phone_number, body_text)
+      "#{calls_phone_id_path}/messages", headers: api_headers, body: permission_request_body(to_phone_number, body_text)
     )
 
     unless response.success?
@@ -30,13 +35,19 @@ module Enterprise::Whatsapp::Providers::WhatsappCloudService
 
   def initiate_call(to_phone_number, sdp_offer)
     response = HTTParty.post(
-      "#{phone_id_path}/calls", headers: api_headers, body: initiate_call_body(to_phone_number, sdp_offer)
+      "#{calls_phone_id_path}/calls", headers: api_headers, body: initiate_call_body(to_phone_number, sdp_offer)
     )
     process_initiate_call_response(response)
   end
 
   private
 
+  def calls_phone_id_path
+    base = ENV.fetch('WHATSAPP_CLOUD_BASE_URL', 'https://graph.facebook.com')
+    version = GlobalConfigService.load('WHATSAPP_API_VERSION', WHATSAPP_CALLING_API_VERSION_FALLBACK)
+    "#{base}/#{version}/#{whatsapp_channel.provider_config['phone_number_id']}"
+  end
+
   def call_action_body(call_id, action, sdp_answer = nil)
     body = { messaging_product: 'whatsapp', call_id: call_id, action: action }
     body[:session] = { sdp: sdp_answer, sdp_type: 'answer' } if sdp_answer
@@ -44,7 +55,7 @@ module Enterprise::Whatsapp::Providers::WhatsappCloudService
   end
 
   def call_api(action_name, body)
-    url = "#{phone_id_path}/calls"
+    url = "#{calls_phone_id_path}/calls"
     Rails.logger.info "[WHATSAPP CALL] #{action_name} POST #{url} body=#{body.except(:session).to_json}"
     response = HTTParty.post(url, headers: api_headers, body: body.to_json)
     Rails.logger.error "[WHATSAPP CALL] #{action_name} failed: status=#{response.code} body=#{response.body}" unless response.success?
@@ -65,7 +76,7 @@ module Enterprise::Whatsapp::Providers::WhatsappCloudService
 
   def initiate_call_body(to_phone_number, sdp_offer)
     {
-      messaging_product: 'whatsapp', to: to_phone_number, type: 'audio',
+      messaging_product: 'whatsapp', to: to_phone_number, action: 'connect',
       session: { sdp: sdp_offer, sdp_type: 'offer' }
     }.to_json
   end
diff --git a/enterprise/app/services/voice/call_message_builder.rb b/enterprise/app/services/voice/call_message_builder.rb
index 8563a8282..11420d60d 100644
--- a/enterprise/app/services/voice/call_message_builder.rb
+++ b/enterprise/app/services/voice/call_message_builder.rb
@@ -7,15 +7,30 @@ class Voice::CallMessageBuilder
     call.message || create_message!
   end
 
+  def update_status!(status:, agent: nil, duration_seconds: nil)
+    message = call.message
+    return unless message
+
+    patch = {
+      'status' => status&.to_s&.tr('_', '-'),
+      'accepted_by' => agent && { 'id' => agent.id, 'name' => agent.name },
+      'duration_seconds' => duration_seconds
+    }.compact
+
+    message.update!(content_attributes: (message.content_attributes || {}).deep_merge('data' => patch))
+    message
+  end
+
   private
 
   attr_reader :call
 
   def create_message!
     params = {
-      content: 'Voice Call',
+      content: I18n.t("conversations.messages.voice_call.#{call.provider}"),
       message_type: call.outgoing? ? 'outgoing' : 'incoming',
-      content_type: 'voice_call'
+      content_type: 'voice_call',
+      content_attributes: { 'data' => build_data_payload }
     }
     Messages::MessageBuilder.new(sender, call.conversation, params).perform
   end
@@ -23,4 +38,15 @@ class Voice::CallMessageBuilder
   def sender
     call.outgoing? ? call.accepted_by_agent : call.contact
   end
+
+  # call_source lets the FE disambiguate WhatsApp vs Twilio without re-fetching the Call.
+  def build_data_payload
+    {
+      'call_id' => call.id,
+      'call_sid' => call.provider_call_id,
+      'call_source' => call.provider,
+      'call_direction' => call.direction_label,
+      'status' => call.display_status
+    }
+  end
 end
diff --git a/enterprise/app/services/voice/inbound_call_builder.rb b/enterprise/app/services/voice/inbound_call_builder.rb
index a2790dda8..7b8b0e684 100644
--- a/enterprise/app/services/voice/inbound_call_builder.rb
+++ b/enterprise/app/services/voice/inbound_call_builder.rb
@@ -1,15 +1,17 @@
 class Voice::InboundCallBuilder
-  attr_reader :account, :inbox, :from_number, :call_sid
+  attr_reader :inbox, :from_number, :call_sid, :provider, :extra_meta
 
-  def self.perform!(account:, inbox:, from_number:, call_sid:)
-    new(account: account, inbox: inbox, from_number: from_number, call_sid: call_sid).perform!
+  def self.perform!(inbox:, from_number:, call_sid:, provider: :twilio, extra_meta: {})
+    new(inbox: inbox, from_number: from_number, call_sid: call_sid,
+        provider: provider, extra_meta: extra_meta).perform!
   end
 
-  def initialize(account:, inbox:, from_number:, call_sid:)
-    @account = account
+  def initialize(inbox:, from_number:, call_sid:, provider: :twilio, extra_meta: {})
     @inbox = inbox
     @from_number = from_number
     @call_sid = call_sid
+    @provider = provider.to_sym
+    @extra_meta = extra_meta || {}
   end
 
   def perform!
@@ -17,8 +19,8 @@ class Voice::InboundCallBuilder
     return existing if existing
 
     ActiveRecord::Base.transaction do
-      contact = ensure_contact!
-      contact_inbox = ensure_contact_inbox!(contact)
+      contact_inbox = ensure_contact_inbox!
+      contact = contact_inbox.contact
       conversation = resolve_conversation!(contact, contact_inbox)
       call = create_call!(contact, conversation)
       message = Voice::CallMessageBuilder.new(call).perform!
@@ -26,15 +28,34 @@ class Voice::InboundCallBuilder
       call
     end
   rescue ActiveRecord::RecordNotUnique
-    # A concurrent Twilio retry won the create race; return what now exists.
+    # A concurrent provider retry won the create race; return what now exists.
     find_existing_call || raise
   end
 
   private
 
+  def account
+    inbox.account
+  end
+
   def find_existing_call
     Call.where(account_id: account.id, inbox_id: inbox.id)
-        .find_by(provider: :twilio, provider_call_id: call_sid)
+        .find_by(provider: provider, provider_call_id: call_sid)
+  end
+
+  # Always look up by (inbox, source_id) first — that pair has a UNIQUE index, so
+  # creating with a colliding source_id under a different contact would raise
+  # RecordNotUnique. Reuse the existing ContactInbox (and its contact) when found.
+  # A concurrent message webhook for the same wa_id can win the (inbox_id, source_id)
+  # race; rescue and re-find so the call path doesn't drop the connect.
+  def ensure_contact_inbox!
+    sid = source_id_for_provider
+    existing = inbox.contact_inboxes.find_by(source_id: sid)
+    return existing if existing
+
+    ContactInbox.create!(contact: ensure_contact!, inbox: inbox, source_id: sid)
+  rescue ActiveRecord::RecordNotUnique
+    inbox.contact_inboxes.find_by!(source_id: sid)
   end
 
   def ensure_contact!
@@ -43,13 +64,14 @@ class Voice::InboundCallBuilder
     end
   end
 
-  def ensure_contact_inbox!(contact)
-    ContactInbox.find_or_create_by!(
-      contact_id: contact.id,
-      inbox_id: inbox.id
-    ) do |record|
-      record.source_id = from_number
-    end
+  # WhatsApp ContactInbox.source_id must be digits-only (the wa_id); Twilio accepts the +.
+  # Run BR/AR-style wa_id normalization (same path messaging uses) so an inbound call
+  # finds the existing ContactInbox instead of forking a new contact/conversation.
+  def source_id_for_provider
+    return from_number unless provider == :whatsapp
+
+    digits = from_number.to_s.delete_prefix('+')
+    Whatsapp::PhoneNumberNormalizationService.new(inbox).normalize_and_find_contact_by_provider(digits, :cloud)
   end
 
   def resolve_conversation!(contact, contact_inbox)
@@ -76,13 +98,14 @@ class Voice::InboundCallBuilder
       inbox: inbox,
       conversation: conversation,
       contact: contact,
-      provider: :twilio,
+      provider: provider,
       direction: :incoming,
       status: 'ringing',
       provider_call_id: call_sid,
-      meta: { 'initiated_at' => Time.zone.now.to_i }
+      meta: { 'initiated_at' => Time.zone.now.to_i }.merge(extra_meta.stringify_keys)
     )
-    call.update!(conference_sid: call.default_conference_sid)
+    # `conference_sid` is a Twilio bridging concept; WhatsApp goes browser↔Meta.
+    call.update!(conference_sid: call.default_conference_sid) if call.twilio?
     call
   end
 end
diff --git a/enterprise/app/services/voice/outbound_call_builder.rb b/enterprise/app/services/voice/outbound_call_builder.rb
index 6c350c5e1..30a74099e 100644
--- a/enterprise/app/services/voice/outbound_call_builder.rb
+++ b/enterprise/app/services/voice/outbound_call_builder.rb
@@ -1,15 +1,16 @@
 class Voice::OutboundCallBuilder
   attr_reader :account, :inbox, :user, :contact
 
-  def self.perform!(account:, inbox:, user:, contact:)
-    new(account: account, inbox: inbox, user: user, contact: contact).perform!
+  def self.perform!(account:, inbox:, user:, contact:, conversation: nil)
+    new(account: account, inbox: inbox, user: user, contact: contact, conversation: conversation).perform!
   end
 
-  def initialize(account:, inbox:, user:, contact:)
+  def initialize(account:, inbox:, user:, contact:, conversation: nil)
     @account = account
     @inbox = inbox
     @user = user
     @contact = contact
+    @existing_conversation = conversation
   end
 
   def perform!
@@ -18,7 +19,7 @@ class Voice::OutboundCallBuilder
 
     ActiveRecord::Base.transaction do
       contact_inbox = ensure_contact_inbox!
-      conversation = create_conversation!(contact_inbox)
+      conversation = @existing_conversation || create_conversation!(contact_inbox)
       call_sid = initiate_call!
       call = create_call!(conversation, call_sid)
       message = Voice::CallMessageBuilder.new(call).perform!
diff --git a/enterprise/app/services/whatsapp/call_permission_reply_service.rb b/enterprise/app/services/whatsapp/call_permission_reply_service.rb
new file mode 100644
index 000000000..f05a6cd2f
--- /dev/null
+++ b/enterprise/app/services/whatsapp/call_permission_reply_service.rb
@@ -0,0 +1,73 @@
+class Whatsapp::CallPermissionReplyService
+  pattr_initialize [:inbox!, :params!]
+
+  def perform
+    return unless inbox.channel.voice_enabled?
+
+    reply_data = extract_reply_data
+    return unless reply_data&.dig(:accepted)
+
+    conversation = find_requesting_conversation(reply_data[:context_id])
+    return unless conversation
+
+    clear_permission_flag(conversation)
+    emit_permission_granted_activity(conversation)
+    broadcast_permission_granted(conversation.contact, conversation)
+  end
+
+  private
+
+  def emit_permission_granted_activity(conversation)
+    content = I18n.t(
+      'conversations.activity.whatsapp_call.permission_granted',
+      contact_name: conversation.contact.name
+    )
+    ::Conversations::ActivityMessageJob.perform_later(
+      conversation,
+      { account_id: conversation.account_id, inbox_id: conversation.inbox_id, message_type: :activity, content: content }
+    )
+  end
+
+  def extract_reply_data
+    message = params.dig(:entry, 0, :changes, 0, :value, :messages, 0)
+    reply = message&.dig(:interactive, :call_permission_reply)
+    return unless reply
+
+    accepted = reply[:response] == 'accept'
+    Rails.logger.info "[WHATSAPP CALL] call_permission_reply from=#{message[:from]} accepted=#{accepted} permanent=#{reply[:is_permanent]}"
+    { from_number: message[:from], accepted: accepted, context_id: message.dig(:context, :id) }
+  end
+
+  # Match the reply to the conversation whose request message it actually points
+  # at (interactive replies carry context.id = our outbound wamid). Recency-based
+  # lookup would broadcast to the wrong thread when a contact has multiple
+  # parallel pending requests.
+  def find_requesting_conversation(context_id)
+    return if context_id.blank?
+
+    inbox.conversations
+         .where.not(status: :resolved)
+         .where("additional_attributes ->> 'call_permission_request_message_id' = ?", context_id)
+         .first
+  end
+
+  def clear_permission_flag(conversation)
+    attrs = (conversation.additional_attributes || {}).except(
+      'call_permission_requested_at', 'call_permission_request_message_id'
+    )
+    conversation.update!(additional_attributes: attrs)
+  end
+
+  def broadcast_permission_granted(contact, conversation)
+    ActionCable.server.broadcast(
+      "account_#{inbox.account_id}",
+      {
+        event: 'voice_call.permission_granted',
+        data: {
+          account_id: inbox.account_id, conversation_id: conversation.id,
+          contact_name: contact.name, contact_phone: contact.phone_number
+        }
+      }
+    )
+  end
+end
diff --git a/enterprise/app/services/whatsapp/call_service.rb b/enterprise/app/services/whatsapp/call_service.rb
new file mode 100644
index 000000000..8a52ea6bc
--- /dev/null
+++ b/enterprise/app/services/whatsapp/call_service.rb
@@ -0,0 +1,110 @@
+class Whatsapp::CallService
+  pattr_initialize [:call!, :agent!, :sdp_answer]
+
+  def accept
+    raise Voice::CallErrors::CallFailed, 'sdp_answer is required' if sdp_answer.blank?
+
+    # All side effects under the lock so a concurrent terminate cannot finalize
+    # the call between status update and the message/conversation/broadcast writes.
+    call.with_lock do
+      transition_to_in_progress!
+      update_message_status('in_progress')
+      update_conversation_call_status(call.display_status)
+      broadcast(:accepted, accepted_by_agent_id: agent.id)
+    end
+    call
+  end
+
+  def reject
+    call.with_lock do
+      next if call.terminal? || call.in_progress?
+
+      invoke_provider!(:reject_call)
+      finalize_call('failed')
+    end
+    call
+  end
+
+  def terminate
+    call.with_lock do
+      next if call.terminal?
+
+      invoke_provider!(:terminate_call)
+      # Compute duration from started_at locally — the webhook arrives after the
+      # call is already terminal and the idempotency guard there bails before it
+      # can fill these fields, so we have to record them here.
+      if call.in_progress?
+        duration = call.started_at ? (Time.current - call.started_at).to_i : nil
+        finalize_call('completed', duration_seconds: duration, end_reason: 'agent_hangup')
+      else
+        # Agent hangs up before contact picks up → no_answer; mirrors the webhook terminate path.
+        finalize_call('no_answer', end_reason: 'agent_hangup')
+      end
+    end
+    call
+  end
+
+  private
+
+  def transition_to_in_progress!
+    # Order matters: in_progress and terminal both make ringing? false, so we have to
+    # branch on in_progress? first to surface the distinct AlreadyAccepted state.
+    raise Voice::CallErrors::AlreadyAccepted, 'Call already accepted by another agent' if call.in_progress?
+    raise Voice::CallErrors::NotRinging, 'Call is not in ringing state' unless call.ringing?
+
+    forward_answer_to_meta!
+    call.update!(status: 'in_progress', accepted_by_agent_id: agent.id, started_at: Time.current,
+                 meta: (call.meta || {}).merge('sdp_answer' => sdp_answer))
+    claim_conversation_for_agent
+  end
+
+  def forward_answer_to_meta!
+    invoke_provider!(:pre_accept_call, sdp_answer)
+    invoke_provider!(:accept_call, sdp_answer)
+  end
+
+  # Take ownership of the conversation if no one holds it; leave assignee alone otherwise (transfer via UI).
+  def claim_conversation_for_agent
+    call.conversation.update!(assignee: agent) if call.conversation.assignee_id.blank?
+  end
+
+  # Raise on Meta failure (bool false or transport error) so callers bail before
+  # finalizing local state — otherwise we'd mark a still-active call as ended
+  # and broadcast voice_call.ended while Meta thinks it's live.
+  def invoke_provider!(method, *)
+    success = call.inbox.channel.provider_service.public_send(method, call.provider_call_id, *)
+    raise Voice::CallErrors::CallFailed, "Meta #{method} failed" unless success
+  rescue Voice::CallErrors::CallFailed
+    raise
+  rescue StandardError => e
+    Rails.logger.error "[WHATSAPP CALL] #{method} failed: #{e.class} #{e.message}"
+    raise Voice::CallErrors::CallFailed, "Meta #{method} failed"
+  end
+
+  def finalize_call(status, **attrs)
+    meta = (call.meta || {}).merge('ended_at' => Time.zone.now.to_i)
+    call.update!(status: status, meta: meta, **attrs)
+    update_message_status(status, duration_seconds: attrs[:duration_seconds])
+    update_conversation_call_status(call.display_status)
+    broadcast(:ended, status: call.display_status)
+  end
+
+  def update_message_status(status, duration_seconds: nil)
+    Voice::CallMessageBuilder.new(call).update_status!(status: status, agent: agent, duration_seconds: duration_seconds)
+  end
+
+  def update_conversation_call_status(status)
+    call.conversation.update!(
+      additional_attributes: (call.conversation.additional_attributes || {}).merge('call_status' => status)
+    )
+  end
+
+  def broadcast(event, **extra)
+    payload = {
+      event: "voice_call.#{event}",
+      data: { id: call.id, call_id: call.provider_call_id, provider: call.provider,
+              conversation_id: call.conversation_id, account_id: call.account_id }.merge(extra)
+    }
+    ActionCable.server.broadcast("account_#{call.account_id}", payload)
+  end
+end
diff --git a/enterprise/app/services/whatsapp/incoming_call_service.rb b/enterprise/app/services/whatsapp/incoming_call_service.rb
new file mode 100644
index 000000000..d290e96c2
--- /dev/null
+++ b/enterprise/app/services/whatsapp/incoming_call_service.rb
@@ -0,0 +1,186 @@
+class Whatsapp::IncomingCallService
+  pattr_initialize [:inbox!, :params!]
+
+  def perform
+    return unless inbox.channel.voice_enabled?
+
+    Array(params[:calls]).each { |c| handle_event(c.with_indifferent_access) }
+    Array(params[:statuses]).each { |s| handle_status(s.with_indifferent_access) }
+  end
+
+  private
+
+  def handle_event(payload)
+    case payload[:event]
+    when 'connect' then handle_connect(payload)
+    when 'terminate' then handle_terminate(payload)
+    else Rails.logger.warn "[WHATSAPP CALL] Unknown call event: #{payload[:event]}"
+    end
+  end
+
+  # Meta's `connect` event for outbound calls fires when the WebRTC tunnel is
+  # up — empirically ~20s before the contact actually answers. The real pickup
+  # is reported as a separate webhook with status=ACCEPTED, and is what
+  # `terminate.start_time` aligns to. Treat ACCEPTED as the pickup transition.
+  def handle_status(payload)
+    return unless payload[:type] == 'call'
+
+    call = Call.whatsapp.find_by(provider_call_id: payload[:id])
+    return unless call
+
+    case payload[:status]
+    when 'ACCEPTED' then mark_outbound_accepted(call, payload)
+    when 'RINGING' then nil # informational
+    else Rails.logger.info "[WHATSAPP CALL] Unhandled call status: #{payload[:status]} for #{payload[:id]}"
+    end
+  end
+
+  # with_lock + reload here serializes against Whatsapp::CallService (agent
+  # actions hold call.with_lock across the Meta API call), so a webhook racing
+  # with a terminate can't overwrite a freshly-finalized terminal status.
+  def mark_outbound_accepted(call, payload)
+    call.with_lock do
+      next unless call.outgoing?
+      next if call.in_progress? || call.terminal?
+
+      started_at = Time.zone.at(payload[:timestamp].to_i) if payload[:timestamp].present?
+      update_call!(call, 'in_progress', started_at: started_at || Time.current)
+      broadcast(call, 'voice_call.outbound_accepted')
+    end
+  end
+
+  def handle_connect(payload)
+    call = Call.whatsapp.find_by(provider_call_id: payload[:id])
+    if call.nil?
+      # Only an `offer` payload is a real inbound caller. An `answer` with no
+      # local row means Meta beat our outbound `Call.create!` (tiny window
+      # between initiate API response and DB insert) — do not mint an inbound
+      # row for it; the next status webhook (or a retry) will find it.
+      return create_inbound_call(payload) if inbound_offer?(payload)
+
+      Rails.logger.warn "[WHATSAPP CALL] Outbound connect for unknown call #{payload[:id]}; skipping"
+      return
+    end
+
+    return accept_outbound_call(call, payload) if call.outgoing?
+
+    Rails.logger.info "[WHATSAPP CALL] Duplicate inbound connect for #{payload[:id]}; ignoring"
+  end
+
+  def inbound_offer?(payload)
+    payload.dig(:session, :sdp_type).to_s.downcase == 'offer'
+  end
+
+  def create_inbound_call(payload)
+    sdp_offer = payload.dig(:session, :sdp)
+    call = Voice::InboundCallBuilder.perform!(
+      inbox: inbox, from_number: "+#{payload[:from]}", call_sid: payload[:id],
+      provider: :whatsapp,
+      extra_meta: { 'sdp_offer' => sdp_offer, 'ice_servers' => Call.default_ice_servers }
+    )
+    update_conversation(call)
+    broadcast_incoming(call, sdp_offer)
+  end
+
+  # `connect` is the WebRTC tunnel-ready signal, not the pickup signal. Apply
+  # Meta's SDP answer so the handshake completes during ringing; the call
+  # stays in `ringing` until status=ACCEPTED arrives. Don't gate on
+  # in_progress: an out-of-order ACCEPTED can flip status before connect is
+  # processed, and dropping the SDP answer would leave the browser without
+  # the data it needs to complete the handshake. Use the stored answer as
+  # the idempotency key instead.
+  def accept_outbound_call(call, payload)
+    call.with_lock do
+      next if call.terminal?
+      next if call.meta&.dig('sdp_answer').present?
+
+      # Pin setup:active so browsers don't renegotiate when Meta echoes actpass.
+      sdp_answer = payload.dig(:session, :sdp)&.gsub('a=setup:actpass', 'a=setup:active')
+      call.update!(meta: (call.meta || {}).merge('sdp_answer' => sdp_answer))
+      broadcast(call, 'voice_call.outbound_connected', sdp_answer: sdp_answer)
+    end
+  end
+
+  def handle_terminate(payload)
+    call = Call.whatsapp.find_by(provider_call_id: payload[:id])
+    if call.nil?
+      # No row yet means either an out-of-order terminate (rare in practice — Meta
+      # delivery is FIFO) or, more dangerously, an outbound terminate landing in
+      # the window between the controller's Meta API call and Call.create!.
+      # Materialising as inbound here would collide with the unique
+      # (provider, provider_call_id) index. Skip; controller commits seal it.
+      Rails.logger.warn "[WHATSAPP CALL] Terminate for unknown call #{payload[:id]}; skipping"
+      return
+    end
+
+    call.with_lock do
+      # Webhook retries can re-deliver terminate after we've already finalized the
+      # call; don't recompute status or a duration=0 retry can flip a completed
+      # short call back to no_answer.
+      next if call.terminal?
+
+      duration = payload[:duration]&.to_i
+      reason = payload[:terminate_reason].to_s
+      status = derive_terminate_status(call, duration, reason)
+      meta = (call.meta || {}).merge('ended_at' => Time.zone.now.to_i)
+      update_call!(call, status, duration_seconds: duration, end_reason: reason, meta: meta)
+      broadcast(call, 'voice_call.ended', status: call.display_status, duration_seconds: call.duration_seconds)
+    end
+  end
+
+  # Provider-reported failures trump the answered/no_answer heuristic. An
+  # in_progress call that Meta later terminates with a failure reason would
+  # otherwise be recorded as 'completed' purely because it had been accepted.
+  FAILURE_REASONS = %w[failed error rejected busy invalid_offer cancelled].freeze
+
+  def derive_terminate_status(call, duration, reason)
+    return 'failed' if FAILURE_REASONS.any? { |r| reason.include?(r) }
+
+    answered?(call, duration) ? 'completed' : 'no_answer'
+  end
+
+  # accepted_by_agent_id is the initiating agent on outbound calls, so it only signals "answered" for inbound.
+  def answered?(call, duration)
+    call.in_progress? || duration.to_i.positive? || (call.incoming? && call.accepted_by_agent_id.present?)
+  end
+
+  def update_call!(call, status, **attrs)
+    call.update!(status: status, **attrs)
+    Voice::CallMessageBuilder.new(call).update_status!(status: status, agent: call.accepted_by_agent,
+                                                       duration_seconds: attrs[:duration_seconds])
+    update_conversation(call)
+  end
+
+  def update_conversation(call)
+    call.conversation.update!(
+      additional_attributes: (call.conversation.additional_attributes || {}).merge(
+        'call_status' => call.display_status, 'call_direction' => call.direction_label
+      )
+    )
+  end
+
+  # Ring the assignee if assigned; otherwise account-wide so any agent can pick up.
+  def broadcast_incoming(call, sdp_offer)
+    contact = call.contact
+    token = call.conversation.assignee&.pubsub_token
+    broadcast(call, 'voice_call.incoming',
+              streams: token ? [token] : account_streams,
+              direction: call.direction_label, inbox_id: call.inbox_id,
+              sdp_offer: sdp_offer, ice_servers: Call.default_ice_servers,
+              caller: { name: contact.name, phone: contact.phone_number, avatar: contact.avatar_url })
+  end
+
+  def broadcast(call, event, streams: account_streams, **extra)
+    payload = { event: event, data: base_payload(call).merge(extra) }
+    streams.each { |s| ActionCable.server.broadcast(s, payload) }
+  end
+
+  def account_streams
+    ["account_#{inbox.account_id}"]
+  end
+
+  def base_payload(call)
+    { account_id: inbox.account_id, id: call.id, call_id: call.provider_call_id,
+      provider: 'whatsapp', conversation_id: call.conversation_id }
+  end
+end
diff --git a/enterprise/app/views/api/v1/accounts/whatsapp_calls/accept.json.jbuilder b/enterprise/app/views/api/v1/accounts/whatsapp_calls/accept.json.jbuilder
new file mode 100644
index 000000000..216e225e3
--- /dev/null
+++ b/enterprise/app/views/api/v1/accounts/whatsapp_calls/accept.json.jbuilder
@@ -0,0 +1 @@
+json.partial! 'api/v1/models/whatsapp_call', call: @call
diff --git a/enterprise/app/views/api/v1/accounts/whatsapp_calls/initiate.json.jbuilder b/enterprise/app/views/api/v1/accounts/whatsapp_calls/initiate.json.jbuilder
new file mode 100644
index 000000000..bdb1fa204
--- /dev/null
+++ b/enterprise/app/views/api/v1/accounts/whatsapp_calls/initiate.json.jbuilder
@@ -0,0 +1,5 @@
+json.status 'calling'
+json.call_id @call.provider_call_id
+json.id @call.id
+json.message_id @message.id
+json.provider 'whatsapp'
diff --git a/enterprise/app/views/api/v1/accounts/whatsapp_calls/reject.json.jbuilder b/enterprise/app/views/api/v1/accounts/whatsapp_calls/reject.json.jbuilder
new file mode 100644
index 000000000..db804fd57
--- /dev/null
+++ b/enterprise/app/views/api/v1/accounts/whatsapp_calls/reject.json.jbuilder
@@ -0,0 +1,2 @@
+json.id @call.id
+json.status @call.display_status
diff --git a/enterprise/app/views/api/v1/accounts/whatsapp_calls/show.json.jbuilder b/enterprise/app/views/api/v1/accounts/whatsapp_calls/show.json.jbuilder
new file mode 100644
index 000000000..216e225e3
--- /dev/null
+++ b/enterprise/app/views/api/v1/accounts/whatsapp_calls/show.json.jbuilder
@@ -0,0 +1 @@
+json.partial! 'api/v1/models/whatsapp_call', call: @call
diff --git a/enterprise/app/views/api/v1/accounts/whatsapp_calls/terminate.json.jbuilder b/enterprise/app/views/api/v1/accounts/whatsapp_calls/terminate.json.jbuilder
new file mode 100644
index 000000000..db804fd57
--- /dev/null
+++ b/enterprise/app/views/api/v1/accounts/whatsapp_calls/terminate.json.jbuilder
@@ -0,0 +1,2 @@
+json.id @call.id
+json.status @call.display_status
diff --git a/enterprise/app/views/api/v1/accounts/whatsapp_calls/upload_recording.json.jbuilder b/enterprise/app/views/api/v1/accounts/whatsapp_calls/upload_recording.json.jbuilder
new file mode 100644
index 000000000..6871426d8
--- /dev/null
+++ b/enterprise/app/views/api/v1/accounts/whatsapp_calls/upload_recording.json.jbuilder
@@ -0,0 +1,2 @@
+json.id @call.id
+json.status @upload_status
diff --git a/enterprise/app/views/api/v1/models/_whatsapp_call.json.jbuilder b/enterprise/app/views/api/v1/models/_whatsapp_call.json.jbuilder
new file mode 100644
index 000000000..0d93696c1
--- /dev/null
+++ b/enterprise/app/views/api/v1/models/_whatsapp_call.json.jbuilder
@@ -0,0 +1,24 @@
+contact = call.conversation&.contact
+
+json.id call.id
+json.call_id call.provider_call_id
+json.provider call.provider
+json.status call.display_status
+json.direction call.direction_label
+json.conversation_id call.conversation_id
+json.inbox_id call.inbox_id
+json.message_id call.message_id
+json.accepted_by_agent_id call.accepted_by_agent_id
+json.elapsed_seconds(call.started_at ? (Time.current - call.started_at).to_i : 0)
+json.sdp_offer call.meta&.dig('sdp_offer')
+json.ice_servers(call.meta&.dig('ice_servers') || Call.default_ice_servers)
+
+if contact
+  json.caller do
+    json.name contact.name
+    json.phone contact.phone_number
+    json.avatar contact.avatar_url
+  end
+else
+  json.caller({})
+end
diff --git a/spec/enterprise/controllers/twilio/voice_controller_spec.rb b/spec/enterprise/controllers/twilio/voice_controller_spec.rb
index 205bb2b59..cde513c86 100644
--- a/spec/enterprise/controllers/twilio/voice_controller_spec.rb
+++ b/spec/enterprise/controllers/twilio/voice_controller_spec.rb
@@ -32,7 +32,6 @@ RSpec.describe 'Twilio::VoiceController', type: :request do
       call.update!(conference_sid: call.default_conference_sid)
 
       expect(Voice::InboundCallBuilder).to receive(:perform!).with(
-        account: account,
         inbox: inbox,
         from_number: from_number,
         call_sid: call_sid
diff --git a/spec/enterprise/services/enterprise/whatsapp/providers/whatsapp_cloud_service_spec.rb b/spec/enterprise/services/enterprise/whatsapp/providers/whatsapp_cloud_service_spec.rb
index 9c90d9ccb..87dc6d70c 100644
--- a/spec/enterprise/services/enterprise/whatsapp/providers/whatsapp_cloud_service_spec.rb
+++ b/spec/enterprise/services/enterprise/whatsapp/providers/whatsapp_cloud_service_spec.rb
@@ -4,8 +4,10 @@ describe Whatsapp::Providers::WhatsappCloudService do
   subject(:service) { described_class.new(whatsapp_channel: whatsapp_channel) }
 
   let(:whatsapp_channel) { create(:channel_whatsapp, provider: 'whatsapp_cloud', validate_provider_config: false, sync_templates: false) }
-  let(:calls_url) { 'https://graph.facebook.com/v13.0/123456789/calls' }
-  let(:messages_url) { 'https://graph.facebook.com/v13.0/123456789/messages' }
+  # Call-flow endpoints use the configured WHATSAPP_API_VERSION (fallback v22.0),
+  # not the OSS v13.0 path locked for legacy /messages compatibility.
+  let(:calls_url) { 'https://graph.facebook.com/v22.0/123456789/calls' }
+  let(:messages_url) { 'https://graph.facebook.com/v22.0/123456789/messages' }
   let(:headers) { { 'Content-Type' => 'application/json' } }
 
   before { stub_request(:get, /message_templates/) }
@@ -40,7 +42,7 @@ describe Whatsapp::Providers::WhatsappCloudService do
   describe '#initiate_call' do
     it 'returns the parsed body on success' do
       stub_request(:post, calls_url)
-        .with(body: { messaging_product: 'whatsapp', to: '15551234567', type: 'audio',
+        .with(body: { messaging_product: 'whatsapp', to: '15551234567', action: 'connect',
                       session: { sdp: 'sdp_offer', sdp_type: 'offer' } }.to_json)
         .to_return(status: 200, body: { messages: [{ id: 'wacall_1' }] }.to_json, headers: headers)
 
diff --git a/spec/enterprise/services/voice/inbound_call_builder_spec.rb b/spec/enterprise/services/voice/inbound_call_builder_spec.rb
index 7af9be8e6..3065875f2 100644
--- a/spec/enterprise/services/voice/inbound_call_builder_spec.rb
+++ b/spec/enterprise/services/voice/inbound_call_builder_spec.rb
@@ -16,7 +16,6 @@ RSpec.describe Voice::InboundCallBuilder do
 
   def perform_builder
     described_class.perform!(
-      account: account,
       inbox: inbox,
       from_number: from_number,
       call_sid: call_sid
@@ -87,6 +86,45 @@ RSpec.describe Voice::InboundCallBuilder do
     end
   end
 
+  context 'when a ContactInbox already exists for the source_id (different contact)' do
+    let!(:original_contact) { create(:contact, account: account, phone_number: '+15550009999') }
+    let!(:original_contact_inbox) do
+      create(:contact_inbox, contact: original_contact, inbox: inbox, source_id: from_number)
+    end
+
+    it 'reuses the existing ContactInbox instead of raising RecordNotUnique' do
+      call = perform_builder
+
+      expect(call.contact).to eq(original_contact)
+      expect(call.conversation.contact_inbox).to eq(original_contact_inbox)
+    end
+  end
+
+  context 'when the WhatsApp wa_id needs Brazil normalization to match an existing ContactInbox' do
+    let(:whatsapp_channel) do
+      create(:channel_whatsapp, account: account, provider: 'whatsapp_cloud',
+                                provider_config: { 'phone_number_id' => '123', 'calling_enabled' => true },
+                                validate_provider_config: false, sync_templates: false)
+    end
+    let(:whatsapp_inbox) { whatsapp_channel.inbox }
+    let!(:stored_contact) { create(:contact, account: account, phone_number: '+5541988887777') }
+    let!(:stored_contact_inbox) do
+      create(:contact_inbox, contact: stored_contact, inbox: whatsapp_inbox, source_id: '5541988887777')
+    end
+
+    it 'reuses the contact via normalized wa_id rather than forking a new ContactInbox' do
+      call = described_class.perform!(
+        inbox: whatsapp_inbox,
+        from_number: '+554188887777',
+        call_sid: 'wacall_br_1',
+        provider: :whatsapp
+      )
+
+      expect(call.contact).to eq(stored_contact)
+      expect(call.conversation.contact_inbox).to eq(stored_contact_inbox)
+    end
+  end
+
   context 'when the inbox has lock_to_single_conversation enabled' do
     let!(:contact) { create(:contact, account: account, phone_number: from_number) }
     let!(:contact_inbox) { create(:contact_inbox, contact: contact, inbox: inbox, source_id: from_number) }
diff --git a/spec/enterprise/services/whatsapp/call_permission_reply_service_spec.rb b/spec/enterprise/services/whatsapp/call_permission_reply_service_spec.rb
new file mode 100644
index 000000000..82a91395c
--- /dev/null
+++ b/spec/enterprise/services/whatsapp/call_permission_reply_service_spec.rb
@@ -0,0 +1,98 @@
+require 'rails_helper'
+
+describe Whatsapp::CallPermissionReplyService do
+  let(:account) { create(:account) }
+  let(:channel) do
+    create(:channel_whatsapp, provider: 'whatsapp_cloud', account: account,
+                              validate_provider_config: false, sync_templates: false)
+  end
+  let(:inbox) { channel.inbox }
+  let(:contact) { create(:contact, account: account, phone_number: '+15550001111') }
+  let!(:contact_inbox) { create(:contact_inbox, contact: contact, inbox: inbox, source_id: '15550001111') }
+  let(:request_wamid) { 'wamid.permission_request_abc' }
+  let!(:conversation) do
+    create(:conversation, account: account, inbox: inbox, contact: contact, contact_inbox: contact_inbox, status: :open,
+                          additional_attributes: {
+                            'call_permission_requested_at' => Time.current.iso8601,
+                            'call_permission_request_message_id' => request_wamid
+                          })
+  end
+
+  before do
+    account.enable_features!('channel_voice')
+    channel.provider_config = channel.provider_config.merge('source' => 'embedded_signup', 'calling_enabled' => true)
+    channel.save!
+  end
+
+  def reply_params(response:, context_id: request_wamid)
+    interactive = { type: 'call_permission_reply',
+                    call_permission_reply: { response: response, is_permanent: false } }
+    message = { from: '15550001111', type: 'interactive', interactive: interactive }
+    message[:context] = { id: context_id } if context_id
+    { entry: [{ changes: [{ value: { messages: [message] } }] }] }
+  end
+
+  it 'clears both permission flags and broadcasts voice_call.permission_granted on accept' do
+    allow(ActionCable.server).to receive(:broadcast)
+
+    described_class.new(inbox: inbox, params: reply_params(response: 'accept')).perform
+
+    attrs = conversation.reload.additional_attributes
+    expect(attrs).not_to include('call_permission_requested_at')
+    expect(attrs).not_to include('call_permission_request_message_id')
+    expect(ActionCable.server).to have_received(:broadcast).with(
+      "account_#{account.id}",
+      hash_including(event: 'voice_call.permission_granted',
+                     data: hash_including(conversation_id: conversation.id))
+    )
+  end
+
+  it 'is a no-op when the contact rejected the request' do
+    allow(ActionCable.server).to receive(:broadcast)
+
+    described_class.new(inbox: inbox, params: reply_params(response: 'reject')).perform
+
+    expect(conversation.reload.additional_attributes).to include('call_permission_requested_at')
+    expect(ActionCable.server).not_to have_received(:broadcast)
+  end
+
+  it 'is a no-op when calling is disabled on the channel' do
+    channel.provider_config = channel.provider_config.merge('calling_enabled' => false)
+    channel.save!
+    allow(ActionCable.server).to receive(:broadcast)
+
+    described_class.new(inbox: inbox, params: reply_params(response: 'accept')).perform
+
+    expect(ActionCable.server).not_to have_received(:broadcast)
+  end
+
+  it 'matches the originating conversation by context.id when the contact has multiple pending requests' do
+    other_request_wamid = 'wamid.permission_request_xyz'
+    other_open = create(:conversation, account: account, inbox: inbox, contact: contact, contact_inbox: contact_inbox,
+                                       status: :open,
+                                       additional_attributes: {
+                                         'call_permission_requested_at' => Time.current.iso8601,
+                                         'call_permission_request_message_id' => other_request_wamid
+                                       })
+    allow(ActionCable.server).to receive(:broadcast)
+
+    described_class.new(inbox: inbox, params: reply_params(response: 'accept', context_id: other_request_wamid)).perform
+
+    # The reply pointed at other_open's request — it should be the cleared one, not `conversation`
+    expect(other_open.reload.additional_attributes).not_to include('call_permission_request_message_id')
+    expect(conversation.reload.additional_attributes).to include('call_permission_request_message_id')
+    expect(ActionCable.server).to have_received(:broadcast).with(
+      "account_#{account.id}",
+      hash_including(data: hash_including(conversation_id: other_open.id))
+    )
+  end
+
+  it 'is a no-op when the reply has no context.id' do
+    allow(ActionCable.server).to receive(:broadcast)
+
+    described_class.new(inbox: inbox, params: reply_params(response: 'accept', context_id: nil)).perform
+
+    expect(conversation.reload.additional_attributes).to include('call_permission_request_message_id')
+    expect(ActionCable.server).not_to have_received(:broadcast)
+  end
+end
diff --git a/spec/enterprise/services/whatsapp/incoming_call_service_spec.rb b/spec/enterprise/services/whatsapp/incoming_call_service_spec.rb
new file mode 100644
index 000000000..53c7974f4
--- /dev/null
+++ b/spec/enterprise/services/whatsapp/incoming_call_service_spec.rb
@@ -0,0 +1,242 @@
+require 'rails_helper'
+
+describe Whatsapp::IncomingCallService do
+  let(:account) { create(:account) }
+  let(:channel) do
+    create(:channel_whatsapp, provider: 'whatsapp_cloud', account: account,
+                              validate_provider_config: false, sync_templates: false)
+  end
+  let(:inbox) { channel.inbox }
+  let(:from_number) { '15550001111' }
+  let(:provider_call_id) { 'wacid_abc' }
+
+  before do
+    account.enable_features!('channel_voice')
+    channel.provider_config = channel.provider_config.merge('source' => 'embedded_signup', 'calling_enabled' => true)
+    channel.save!
+  end
+
+  def call_payload(event:, **extra)
+    { calls: [{ id: provider_call_id, from: from_number, event: event, **extra }] }
+  end
+
+  context 'when calling is disabled on the channel' do
+    it 'is a no-op' do
+      channel.provider_config = channel.provider_config.merge('calling_enabled' => false)
+      channel.save!
+
+      expect { described_class.new(inbox: inbox, params: call_payload(event: 'connect')).perform }
+        .not_to change(Call, :count)
+    end
+  end
+
+  describe 'inbound connect' do
+    let(:sdp_offer) { "v=0\r\n...sdp..." }
+
+    it 'creates the Call + Conversation + voice_call message and broadcasts voice_call.incoming' do
+      allow(ActionCable.server).to receive(:broadcast)
+
+      params = call_payload(event: 'connect', session: { sdp: sdp_offer, sdp_type: 'offer' })
+      expect { described_class.new(inbox: inbox, params: params).perform }
+        .to change(Call, :count).by(1).and change(Conversation, :count).by(1)
+
+      call = Call.last
+      expect(call).to have_attributes(provider: 'whatsapp', direction: 'incoming', status: 'ringing',
+                                      provider_call_id: provider_call_id)
+      expect(call.meta['sdp_offer']).to eq(sdp_offer)
+      expect(ActionCable.server).to have_received(:broadcast).with(
+        "account_#{account.id}",
+        hash_including(event: 'voice_call.incoming', data: hash_including(sdp_offer: sdp_offer))
+      )
+    end
+  end
+
+  describe 'outbound connect (existing call)' do
+    let!(:call) do
+      conversation = create(:conversation, account: account, inbox: inbox)
+      create(:call, account: account, inbox: inbox, conversation: conversation, contact: conversation.contact,
+                    provider: :whatsapp, direction: :outgoing, status: 'ringing', provider_call_id: provider_call_id)
+    end
+
+    it 'stores the SDP answer and broadcasts voice_call.outbound_connected without flipping to in_progress' do
+      allow(ActionCable.server).to receive(:broadcast)
+      sdp_answer = "v=0\r\na=setup:actpass\r\n"
+
+      params = call_payload(event: 'connect', session: { sdp: sdp_answer, sdp_type: 'answer' })
+      described_class.new(inbox: inbox, params: params).perform
+
+      # connect only completes the SDP handshake; pickup is reported separately as status=ACCEPTED.
+      expect(call.reload).to have_attributes(status: 'ringing', started_at: nil)
+      expect(call.meta['sdp_answer']).to include('a=setup:active')
+      expect(ActionCable.server).to have_received(:broadcast).with(
+        "account_#{account.id}",
+        hash_including(event: 'voice_call.outbound_connected')
+      )
+    end
+  end
+
+  describe 'terminate' do
+    let!(:call) do
+      conversation = create(:conversation, account: account, inbox: inbox)
+      create(:call, account: account, inbox: inbox, conversation: conversation, contact: conversation.contact,
+                    provider: :whatsapp, direction: :incoming, status: 'in_progress', provider_call_id: provider_call_id)
+    end
+
+    it 'marks the call completed when the call had been answered' do
+      allow(ActionCable.server).to receive(:broadcast)
+
+      params = call_payload(event: 'terminate', duration: 42, terminate_reason: 'completed_normally')
+      described_class.new(inbox: inbox, params: params).perform
+
+      expect(call.reload).to have_attributes(status: 'completed', duration_seconds: 42, end_reason: 'completed_normally')
+      expect(call.ended_at).to be_present
+      expect(ActionCable.server).to have_received(:broadcast).with(
+        "account_#{account.id}",
+        hash_including(event: 'voice_call.ended', data: hash_including(status: 'completed'))
+      )
+    end
+
+    it 'marks unanswered ringing calls as no_answer' do
+      call.update!(status: 'ringing')
+      params = call_payload(event: 'terminate', duration: 0, terminate_reason: 'no_answer')
+
+      described_class.new(inbox: inbox, params: params).perform
+      expect(call.reload.status).to eq('no_answer')
+    end
+
+    it 'records the call as failed when Meta reports a failure reason for an in_progress call' do
+      allow(ActionCable.server).to receive(:broadcast)
+
+      params = call_payload(event: 'terminate', duration: 12, terminate_reason: 'failed')
+      described_class.new(inbox: inbox, params: params).perform
+
+      expect(call.reload).to have_attributes(status: 'failed', duration_seconds: 12, end_reason: 'failed')
+    end
+
+    it 'is a no-op when the call is already terminal so retries cannot flip a completed call to no_answer' do
+      call.update!(status: 'completed', duration_seconds: 5, direction: :outgoing, accepted_by_agent: nil)
+      allow(ActionCable.server).to receive(:broadcast)
+      params = call_payload(event: 'terminate', duration: 0, terminate_reason: 'completed_normally')
+
+      described_class.new(inbox: inbox, params: params).perform
+
+      expect(call.reload).to have_attributes(status: 'completed', duration_seconds: 5)
+      expect(ActionCable.server).not_to have_received(:broadcast)
+    end
+  end
+
+  describe 'terminate with no local row yet' do
+    it 'logs and skips instead of materialising an inbound missed-call row' do
+      allow(Rails.logger).to receive(:warn)
+      allow(ActionCable.server).to receive(:broadcast)
+
+      params = call_payload(event: 'terminate', duration: 0, terminate_reason: 'no_answer')
+
+      expect { described_class.new(inbox: inbox, params: params).perform }
+        .not_to change(Call, :count)
+      expect(Rails.logger).to have_received(:warn).with(/Terminate for unknown call/)
+      expect(ActionCable.server).not_to have_received(:broadcast)
+    end
+  end
+
+  describe 'outbound connect with no local row yet' do
+    it 'does not mint an inbound call when sdp_type is answer' do
+      allow(Rails.logger).to receive(:warn)
+      allow(ActionCable.server).to receive(:broadcast)
+
+      params = call_payload(event: 'connect', session: { sdp: 'sdp_answer', sdp_type: 'answer' })
+
+      expect { described_class.new(inbox: inbox, params: params).perform }
+        .not_to change(Call, :count)
+      expect(Rails.logger).to have_received(:warn).with(/Outbound connect for unknown call/)
+      expect(ActionCable.server).not_to have_received(:broadcast)
+    end
+  end
+
+  describe 'duplicate inbound connect' do
+    let!(:call) do
+      conversation = create(:conversation, account: account, inbox: inbox)
+      create(:call, account: account, inbox: inbox, conversation: conversation, contact: conversation.contact,
+                    provider: :whatsapp, direction: :incoming, status: 'ringing', provider_call_id: provider_call_id)
+    end
+
+    it 'logs and ignores rather than treating it as outbound' do
+      allow(Rails.logger).to receive(:info)
+      allow(ActionCable.server).to receive(:broadcast)
+      params = call_payload(event: 'connect', session: { sdp: 'sdp_x', sdp_type: 'offer' })
+
+      described_class.new(inbox: inbox, params: params).perform
+
+      expect(call.reload).to have_attributes(status: 'ringing')
+      expect(Rails.logger).to have_received(:info).with(/Duplicate inbound connect/)
+      expect(ActionCable.server).not_to have_received(:broadcast)
+    end
+  end
+
+  describe 'connect arriving after terminal status' do
+    let!(:call) do
+      conversation = create(:conversation, account: account, inbox: inbox)
+      create(:call, account: account, inbox: inbox, conversation: conversation, contact: conversation.contact,
+                    provider: :whatsapp, direction: :outgoing, status: 'completed', provider_call_id: provider_call_id)
+    end
+
+    it 'does not reopen a completed outbound call' do
+      allow(ActionCable.server).to receive(:broadcast)
+      params = call_payload(event: 'connect', session: { sdp: 'late_sdp', sdp_type: 'answer' })
+
+      described_class.new(inbox: inbox, params: params).perform
+
+      expect(call.reload.status).to eq('completed')
+      expect(ActionCable.server).not_to have_received(:broadcast)
+    end
+  end
+
+  describe 'unanswered outbound call terminate' do
+    let!(:agent) { create(:user, account: account) }
+    let!(:call) do
+      conversation = create(:conversation, account: account, inbox: inbox)
+      create(:call, account: account, inbox: inbox, conversation: conversation, contact: conversation.contact,
+                    provider: :whatsapp, direction: :outgoing, status: 'ringing',
+                    accepted_by_agent: agent, provider_call_id: provider_call_id)
+    end
+
+    it 'marks the call as no_answer even though accepted_by_agent_id is set' do
+      allow(ActionCable.server).to receive(:broadcast)
+      params = call_payload(event: 'terminate', duration: 0, terminate_reason: 'no_answer')
+
+      described_class.new(inbox: inbox, params: params).perform
+
+      expect(call.reload.status).to eq('no_answer')
+      expect(ActionCable.server).to have_received(:broadcast).with(
+        "account_#{account.id}",
+        hash_including(event: 'voice_call.ended', data: hash_including(status: 'no-answer'))
+      )
+    end
+  end
+
+  describe 'unknown event' do
+    it 'logs a warning and does not raise' do
+      allow(Rails.logger).to receive(:warn)
+      params = call_payload(event: 'mystery')
+
+      expect { described_class.new(inbox: inbox, params: params).perform }.not_to raise_error
+      expect(Rails.logger).to have_received(:warn).with(/Unknown call event: mystery/)
+    end
+  end
+
+  describe 'multiple calls in one webhook payload' do
+    it 'processes every call in the array' do
+      allow(ActionCable.server).to receive(:broadcast)
+
+      params = {
+        calls: [
+          { id: 'wacid_a', from: from_number, event: 'connect', session: { sdp: 'sdp_a', sdp_type: 'offer' } },
+          { id: 'wacid_b', from: '15550002222', event: 'connect', session: { sdp: 'sdp_b', sdp_type: 'offer' } }
+        ]
+      }
+
+      expect { described_class.new(inbox: inbox, params: params).perform }.to change(Call, :count).by(2)
+      expect(Call.where(provider_call_id: %w[wacid_a wacid_b]).pluck(:provider_call_id)).to contain_exactly('wacid_a', 'wacid_b')
+    end
+  end
+end
diff --git a/spec/services/telegram/send_attachments_service_spec.rb b/spec/services/telegram/send_attachments_service_spec.rb
index f9c625780..739dd6f73 100644
--- a/spec/services/telegram/send_attachments_service_spec.rb
+++ b/spec/services/telegram/send_attachments_service_spec.rb
@@ -41,7 +41,9 @@ RSpec.describe Telegram::SendAttachmentsService do
     end
 
     context 'when this is business chat' do
-      before { allow(channel).to receive(:business_connection_id).and_return('eooW3KF5WB5HxTD7T826') }
+      before do
+        message.conversation.update!(additional_attributes: { 'business_connection_id' => 'eooW3KF5WB5HxTD7T826' })
+      end
 
       it 'sends all types of attachments in seperate groups and returns the last successful message ID from the batch' do
         attach_files(message)
diff --git a/spec/services/whatsapp/facebook_api_client_spec.rb b/spec/services/whatsapp/facebook_api_client_spec.rb
index 74fb2f6e2..a33999bee 100644
--- a/spec/services/whatsapp/facebook_api_client_spec.rb
+++ b/spec/services/whatsapp/facebook_api_client_spec.rb
@@ -177,7 +177,7 @@ describe Whatsapp::FacebookApiClient do
           .with(
             headers: { 'Authorization' => "Bearer #{access_token}", 'Content-Type' => 'application/json' },
             body: { override_callback_uri: callback_url, verify_token: verify_token,
-                    subscribed_fields: %w[messages smb_message_echoes] }.to_json
+                    subscribed_fields: %w[messages smb_message_echoes calls] }.to_json
           )
           .to_return(
             status: 200,
@@ -224,7 +224,7 @@ describe Whatsapp::FacebookApiClient do
           .with(
             headers: { 'Authorization' => "Bearer #{access_token}", 'Content-Type' => 'application/json' },
             body: { override_callback_uri: callback_url, verify_token: verify_token,
-                    subscribed_fields: %w[messages smb_message_echoes] }.to_json
+                    subscribed_fields: %w[messages smb_message_echoes calls] }.to_json
           )
           .to_return(status: 400, body: { error: 'Webhook callback override failed' }.to_json)
       end

From 3df827c93156bd147e92b0ca992e33ba1989a51e Mon Sep 17 00:00:00 2001
From: Sivin Varghese <64252451+iamsivin@users.noreply.github.com>
Date: Tue, 12 May 2026 11:27:30 +0530
Subject: [PATCH 17/20] chore: update Captain documents filter UI (#14429)

---
 .../captain/assistant/DocumentFilter.vue      | 17 +----
 .../captain/assistant/DocumentFiltersBar.vue  | 68 ++++++-------------
 .../dashboard/captain/documents/Index.vue     | 31 ++++++++-
 3 files changed, 53 insertions(+), 63 deletions(-)

diff --git a/app/javascript/dashboard/components-next/captain/assistant/DocumentFilter.vue b/app/javascript/dashboard/components-next/captain/assistant/DocumentFilter.vue
index 63b06daa7..57d1f32b0 100644
--- a/app/javascript/dashboard/components-next/captain/assistant/DocumentFilter.vue
+++ b/app/javascript/dashboard/components-next/captain/assistant/DocumentFilter.vue
@@ -1,6 +1,5 @@
 <script setup>
 import { computed, ref } from 'vue';
-import { debounce } from '@chatwoot/utils';
 
 import DocumentFiltersBar from 'dashboard/components-next/captain/assistant/DocumentFiltersBar.vue';
 
@@ -9,13 +8,9 @@ const emit = defineEmits(['change']);
 const source = ref('all');
 const status = ref(null);
 const sort = ref('recently_updated');
-const query = ref('');
 
 const hasActiveFilters = computed(
-  () =>
-    source.value !== 'all' ||
-    Boolean(status.value) ||
-    Boolean(query.value.trim())
+  () => source.value !== 'all' || Boolean(status.value)
 );
 
 const buildParams = (page = 1) => {
@@ -23,12 +18,10 @@ const buildParams = (page = 1) => {
   if (source.value !== 'all') params.source = source.value;
   if (status.value) params.filter = status.value;
   if (sort.value) params.sort = sort.value;
-  if (query.value.trim()) params.searchKey = query.value.trim();
   return params;
 };
 
 const emitChange = () => emit('change');
-const debouncedEmitChange = debounce(emitChange, 300);
 
 const handleSourceSelect = sourceKey => {
   source.value = sourceKey;
@@ -47,16 +40,10 @@ const handleSortSelect = sortKey => {
   emitChange();
 };
 
-const handleSearch = value => {
-  query.value = value;
-  debouncedEmitChange();
-};
-
 const reset = () => {
   source.value = 'all';
   status.value = null;
   sort.value = 'recently_updated';
-  query.value = '';
 };
 
 defineExpose({ buildParams, reset, hasActiveFilters });
@@ -67,10 +54,8 @@ defineExpose({ buildParams, reset, hasActiveFilters });
     :active-source-filter="source"
     :active-status-filter="status"
     :active-sort="sort"
-    :search-query="query"
     @select-source="handleSourceSelect"
     @select-status="handleStatusSelect"
     @select-sort="handleSortSelect"
-    @search="handleSearch"
   />
 </template>
diff --git a/app/javascript/dashboard/components-next/captain/assistant/DocumentFiltersBar.vue b/app/javascript/dashboard/components-next/captain/assistant/DocumentFiltersBar.vue
index 30d44b42e..b0a5b9dd0 100644
--- a/app/javascript/dashboard/components-next/captain/assistant/DocumentFiltersBar.vue
+++ b/app/javascript/dashboard/components-next/captain/assistant/DocumentFiltersBar.vue
@@ -1,9 +1,9 @@
 <script setup>
 import { computed, ref } from 'vue';
 import { useI18n } from 'vue-i18n';
+import { vOnClickOutside } from '@vueuse/components';
 
 import Icon from 'dashboard/components-next/icon/Icon.vue';
-import Input from 'dashboard/components-next/input/Input.vue';
 import Button from 'dashboard/components-next/button/Button.vue';
 import DropdownMenu from 'dashboard/components-next/dropdown-menu/DropdownMenu.vue';
 
@@ -11,15 +11,9 @@ const props = defineProps({
   activeSourceFilter: { type: String, default: 'all' },
   activeStatusFilter: { type: String, default: null },
   activeSort: { type: String, default: 'recently_updated' },
-  searchQuery: { type: String, default: '' },
 });
 
-const emit = defineEmits([
-  'selectSource',
-  'selectStatus',
-  'selectSort',
-  'search',
-]);
+const emit = defineEmits(['selectSource', 'selectStatus', 'selectSort']);
 
 const { t } = useI18n();
 
@@ -121,45 +115,27 @@ const handleMenuAction = ({ action, value }) => {
 
 <template>
   <div
-    v-on-clickaway="closeMenu"
-    class="flex flex-col gap-3 w-full lg:flex-row lg:items-center lg:justify-between"
+    v-on-click-outside="closeMenu"
+    class="inline-flex flex-wrap items-center gap-2 pt-2 w-fit"
   >
-    <div class="flex flex-wrap items-center gap-2">
-      <div v-for="menu in filterMenus" :key="menu.key" class="relative">
-        <Button
-          :icon="menu.selected.icon"
-          slate
-          outline
-          :class="{ 'bg-n-slate-9/10': openMenu === menu.key }"
-          @click="toggleMenu(menu.key)"
-        >
-          <span class="min-w-0 truncate">{{ menu.selected.label }}</span>
-          <Icon icon="i-lucide-chevron-down" class="shrink-0 size-4" />
-        </Button>
-        <DropdownMenu
-          v-if="openMenu === menu.key"
-          :menu-items="menu.items"
-          :class="menu.dropdownClass"
-          class="top-full mt-2 ltr:left-0 rtl:right-0"
-          @action="handleMenuAction"
-        />
-      </div>
+    <div v-for="menu in filterMenus" :key="menu.key" class="relative">
+      <Button
+        :icon="menu.selected.icon"
+        slate
+        size="sm"
+        :class="{ 'bg-n-slate-9/10': openMenu === menu.key }"
+        @click="toggleMenu(menu.key)"
+      >
+        <span class="min-w-0 truncate">{{ menu.selected.label }}</span>
+        <Icon icon="i-lucide-chevron-down" class="shrink-0 size-4" />
+      </Button>
+      <DropdownMenu
+        v-if="openMenu === menu.key"
+        :menu-items="menu.items"
+        :class="menu.dropdownClass"
+        class="top-full mt-2 ltr:left-0 rtl:right-0"
+        @action="handleMenuAction"
+      />
     </div>
-
-    <Input
-      :model-value="searchQuery"
-      type="search"
-      :placeholder="t('CAPTAIN.DOCUMENTS.FILTERS.SEARCH_PLACEHOLDER')"
-      :custom-input-class="['ltr:!pl-9 rtl:!pr-9']"
-      class="w-full lg:max-w-72"
-      @input="emit('search', $event.target.value)"
-    >
-      <template #prefix>
-        <Icon
-          icon="i-lucide-search"
-          class="absolute size-4 text-n-slate-11 top-1/2 -translate-y-1/2 ltr:left-3 rtl:right-3"
-        />
-      </template>
-    </Input>
   </div>
 </template>
diff --git a/app/javascript/dashboard/routes/dashboard/captain/documents/Index.vue b/app/javascript/dashboard/routes/dashboard/captain/documents/Index.vue
index e13207416..87c04aefe 100644
--- a/app/javascript/dashboard/routes/dashboard/captain/documents/Index.vue
+++ b/app/javascript/dashboard/routes/dashboard/captain/documents/Index.vue
@@ -7,11 +7,13 @@ import { FEATURE_FLAGS } from 'dashboard/featureFlags';
 import { useAccount } from 'dashboard/composables/useAccount';
 import { useAlert } from 'dashboard/composables';
 import { usePolicy } from 'dashboard/composables/usePolicy';
+import { debounce } from '@chatwoot/utils';
 
 import DeleteDialog from 'dashboard/components-next/captain/pageComponents/DeleteDialog.vue';
 import DocumentCard from 'dashboard/components-next/captain/assistant/DocumentCard.vue';
 import DocumentFilter from 'dashboard/components-next/captain/assistant/DocumentFilter.vue';
 import DocumentBulkActions from 'dashboard/components-next/captain/assistant/DocumentBulkActions.vue';
+import Input from 'dashboard/components-next/input/Input.vue';
 import Policy from 'dashboard/components/policy.vue';
 import PageLayout from 'dashboard/components-next/captain/PageLayout.vue';
 import CaptainPaywall from 'dashboard/components-next/captain/pageComponents/Paywall.vue';
@@ -73,6 +75,7 @@ const handleCreateDialogClose = () => {
 
 const documentFilter = ref(null);
 const syncIntervalHours = ref(null);
+const searchQuery = ref('');
 
 const currentAssistantId = () =>
   Number.isFinite(selectedAssistantId.value) ? selectedAssistantId.value : null;
@@ -84,6 +87,8 @@ const buildDocumentFilterParams = (page = 1) => {
   };
   const assistantId = currentAssistantId();
   if (assistantId) filterParams.assistantId = assistantId;
+  const trimmedQuery = searchQuery.value.trim();
+  if (trimmedQuery) filterParams.searchKey = trimmedQuery;
   return filterParams;
 };
 
@@ -154,6 +159,11 @@ const onFiltersChanged = () => {
   fetchDocuments(1);
 };
 
+const debouncedSearch = debounce(() => {
+  bulkSelectedIds.value = new Set();
+  fetchDocuments(1);
+}, 300);
+
 const syncPollStartedAt = ref(null);
 
 const hasDocumentsSyncing = computed(() =>
@@ -284,13 +294,16 @@ const onCreateSuccess = () => {
 };
 
 const hasActiveDocumentFilters = computed(
-  () => documentFilter.value?.hasActiveFilters ?? false
+  () =>
+    (documentFilter.value?.hasActiveFilters ?? false) ||
+    Boolean(searchQuery.value.trim())
 );
 
 watch(
   selectedAssistantId,
   async () => {
     documentFilter.value?.reset();
+    searchQuery.value = '';
     bulkSelectedIds.value = new Set();
     syncIntervalHours.value = null;
     stopSyncPolling();
@@ -320,6 +333,22 @@ onUnmounted(() => {
     @update:current-page="onPageChange"
     @click="handleCreateDocument"
   >
+    <template #search>
+      <div
+        v-if="bulkSelectedIds.size === 0"
+        class="flex gap-3 justify-between w-full items-center"
+      >
+        <Input
+          v-model="searchQuery"
+          :placeholder="$t('CAPTAIN.DOCUMENTS.FILTERS.SEARCH_PLACEHOLDER')"
+          class="max-w-64 min-w-0 w-full"
+          size="sm"
+          type="search"
+          @input="debouncedSearch"
+        />
+      </div>
+    </template>
+
     <template #subHeader>
       <Policy :permissions="['administrator']">
         <DocumentBulkActions

From 6c67eb9ba0e526da8bc6f99fc530dcab340967bc Mon Sep 17 00:00:00 2001
From: Pranav <pranav@chatwoot.com>
Date: Mon, 11 May 2026 23:57:29 -0700
Subject: [PATCH 18/20] fix(notifications): Respect conversation access when
 notifying agents (#14412)

Agents with limited custom roles were receiving notifications (creation,
assignment, mentions, new messages, SLA) for conversations they couldn't
actually open. For example, an agent whose custom role only grants
`conversation_unassigned_manage` was getting notified about
conversations assigned to other agents.

Notifications now go through the same `ConversationPolicy#show?` check
that gates the conversation view itself, so an agent only gets notified
for conversations they're permitted to see. Administrators and agents
without custom roles are unaffected.

---------

Co-authored-by: Sojan Jose <sojan@pepalo.com>
Co-authored-by: Muhsin Keloth <muhsinkeramam@gmail.com>
---
 app/builders/notification_builder.rb          |  15 +++
 app/services/messages/mention_service.rb      |   2 +-
 spec/builders/notification_builder_spec.rb    |  65 ++++++++++-
 .../builders/notification_builder_spec.rb     | 106 ++++++++++++++++++
 spec/enterprise/models/sla_event_spec.rb      |   1 +
 .../services/messages/mention_service_spec.rb |  20 ++++
 6 files changed, 207 insertions(+), 2 deletions(-)
 create mode 100644 spec/enterprise/builders/notification_builder_spec.rb

diff --git a/app/builders/notification_builder.rb b/app/builders/notification_builder.rb
index d5461eb4e..e09e44d50 100644
--- a/app/builders/notification_builder.rb
+++ b/app/builders/notification_builder.rb
@@ -27,6 +27,8 @@ class NotificationBuilder
     return if notification_type == 'conversation_creation' && !user_subscribed_to_notification?
     # skip notifications for blocked conversations except for user mentions
     return if primary_actor.contact.blocked? && notification_type != 'conversation_mention'
+    # respect conversation access (inbox/team membership and custom-role permissions)
+    return unless user_can_access_conversation?
 
     user.notifications.create!(
       notification_type: notification_type,
@@ -36,4 +38,17 @@ class NotificationBuilder
       secondary_actor: secondary_actor || current_user
     )
   end
+
+  def user_can_access_conversation?
+    conversation = primary_actor.is_a?(Conversation) ? primary_actor : primary_actor.try(:conversation)
+    return true if conversation.blank?
+
+    account_user = AccountUser.find_by(account_id: account.id, user_id: user.id)
+    return false if account_user.blank?
+
+    ConversationPolicy.new(
+      { user: user, account: account, account_user: account_user },
+      conversation
+    ).show?
+  end
 end
diff --git a/app/services/messages/mention_service.rb b/app/services/messages/mention_service.rb
index 8171f8168..1d42498a9 100644
--- a/app/services/messages/mention_service.rb
+++ b/app/services/messages/mention_service.rb
@@ -8,8 +8,8 @@ class Messages::MentionService
     return if validated_mentioned_ids.blank?
 
     Conversations::UserMentionJob.perform_later(validated_mentioned_ids, message.conversation.id, message.account.id)
-    generate_notifications_for_mentions(validated_mentioned_ids)
     add_mentioned_users_as_participants(validated_mentioned_ids)
+    generate_notifications_for_mentions(validated_mentioned_ids)
   end
 
   private
diff --git a/spec/builders/notification_builder_spec.rb b/spec/builders/notification_builder_spec.rb
index 1b28f6adf..f191f0485 100644
--- a/spec/builders/notification_builder_spec.rb
+++ b/spec/builders/notification_builder_spec.rb
@@ -6,9 +6,11 @@ describe NotificationBuilder do
   describe '#perform' do
     let!(:account) { create(:account) }
     let!(:user) { create(:user, account: account) }
-    let!(:primary_actor) { create(:conversation, account: account) }
+    let!(:inbox) { create(:inbox, account: account) }
+    let!(:primary_actor) { create(:conversation, account: account, inbox: inbox) }
 
     before do
+      create(:inbox_member, user: user, inbox: inbox)
       notification_setting = user.notification_settings.find_by(account_id: account.id)
       notification_setting.selected_email_flags = [:email_conversation_creation]
       notification_setting.selected_push_flags = [:push_conversation_creation]
@@ -97,5 +99,66 @@ describe NotificationBuilder do
         ).perform
       end.to change { user.notifications.count }.by(1)
     end
+
+    context 'when the user does not have access to the conversation' do
+      let!(:outsider) { create(:user, account: account) }
+
+      it 'does not create a notification for an agent without inbox or team access' do
+        expect do
+          described_class.new(
+            notification_type: 'conversation_creation',
+            user: outsider,
+            account: account,
+            primary_actor: primary_actor
+          ).perform
+        end.not_to(change { outsider.notifications.count })
+      end
+
+      it 'still creates a notification for administrators regardless of inbox membership' do
+        admin = create(:user, account: account, role: :administrator)
+        admin_setting = admin.notification_settings.find_by(account_id: account.id)
+        admin_setting.selected_email_flags = [:email_conversation_creation]
+        admin_setting.selected_push_flags = [:push_conversation_creation]
+        admin_setting.save!
+
+        expect do
+          described_class.new(
+            notification_type: 'conversation_creation',
+            user: admin,
+            account: account,
+            primary_actor: primary_actor
+          ).perform
+        end.to change { admin.notifications.count }.by(1)
+      end
+
+      it 'does not create a notification when the user is not part of the account' do
+        unrelated_user = create(:user)
+
+        expect do
+          described_class.new(
+            notification_type: 'conversation_creation',
+            user: unrelated_user,
+            account: account,
+            primary_actor: primary_actor
+          ).perform
+        end.not_to(change { unrelated_user.notifications.count })
+      end
+
+      it 'derives the conversation from a message primary_actor' do
+        outsider_inbox = create(:inbox, account: account)
+        message = create(:message, account: account, inbox: outsider_inbox,
+                                   conversation: create(:conversation, account: account, inbox: outsider_inbox))
+
+        expect do
+          described_class.new(
+            notification_type: 'conversation_mention',
+            user: outsider,
+            account: account,
+            primary_actor: message.conversation,
+            secondary_actor: message
+          ).perform
+        end.not_to(change { outsider.notifications.count })
+      end
+    end
   end
 end
diff --git a/spec/enterprise/builders/notification_builder_spec.rb b/spec/enterprise/builders/notification_builder_spec.rb
new file mode 100644
index 000000000..e25034f0c
--- /dev/null
+++ b/spec/enterprise/builders/notification_builder_spec.rb
@@ -0,0 +1,106 @@
+require 'rails_helper'
+
+describe NotificationBuilder do
+  describe '#perform with custom role permissions' do
+    let!(:account) { create(:account) }
+    let!(:agent) { create(:user, account: account, role: :agent) }
+    let!(:inbox) { create(:inbox, account: account) }
+    let!(:account_user) { agent.account_users.find_by(account: account) }
+
+    before do
+      create(:inbox_member, user: agent, inbox: inbox)
+      notification_setting = agent.notification_settings.find_by(account_id: account.id)
+      notification_setting.selected_email_flags = [:email_conversation_creation]
+      notification_setting.selected_push_flags = [:push_conversation_creation]
+      notification_setting.save!
+    end
+
+    def build_notification(conversation, type: 'conversation_creation')
+      described_class.new(
+        notification_type: type,
+        user: agent,
+        account: account,
+        primary_actor: conversation
+      ).perform
+    end
+
+    context 'when the agent has conversation_manage permission' do
+      before do
+        custom_role = create(:custom_role, account: account, permissions: ['conversation_manage'])
+        account_user.update!(custom_role: custom_role)
+      end
+
+      it 'creates a notification for any inbox conversation' do
+        conversation = create(:conversation, account: account, inbox: inbox)
+
+        expect { build_notification(conversation) }.to change { agent.notifications.count }.by(1)
+      end
+    end
+
+    context 'when the agent has conversation_unassigned_manage permission' do
+      before do
+        custom_role = create(:custom_role, account: account, permissions: ['conversation_unassigned_manage'])
+        account_user.update!(custom_role: custom_role)
+      end
+
+      it 'creates a notification for unassigned conversations' do
+        conversation = create(:conversation, account: account, inbox: inbox, assignee: nil)
+
+        expect { build_notification(conversation) }.to change { agent.notifications.count }.by(1)
+      end
+
+      it 'creates a notification for conversations assigned to the agent' do
+        conversation = create(:conversation, account: account, inbox: inbox, assignee: agent)
+
+        expect { build_notification(conversation) }.to change { agent.notifications.count }.by(1)
+      end
+
+      it 'does not create a notification for conversations assigned to someone else' do
+        other_agent = create(:user, account: account, role: :agent)
+        create(:inbox_member, user: other_agent, inbox: inbox)
+        conversation = create(:conversation, account: account, inbox: inbox, assignee: other_agent)
+
+        expect { build_notification(conversation) }.not_to(change { agent.notifications.count })
+      end
+    end
+
+    context 'when the agent has conversation_participating_manage permission' do
+      before do
+        custom_role = create(:custom_role, account: account, permissions: ['conversation_participating_manage'])
+        account_user.update!(custom_role: custom_role)
+      end
+
+      it 'creates a notification for conversations assigned to the agent' do
+        conversation = create(:conversation, account: account, inbox: inbox, assignee: agent)
+
+        expect { build_notification(conversation) }.to change { agent.notifications.count }.by(1)
+      end
+
+      it 'creates a notification for conversations the agent participates in' do
+        conversation = create(:conversation, account: account, inbox: inbox, assignee: nil)
+        create(:conversation_participant, conversation: conversation, account: account, user: agent)
+
+        expect { build_notification(conversation) }.to change { agent.notifications.count }.by(1)
+      end
+
+      it 'does not create a notification for unassigned conversations the agent does not participate in' do
+        conversation = create(:conversation, account: account, inbox: inbox, assignee: nil)
+
+        expect { build_notification(conversation) }.not_to(change { agent.notifications.count })
+      end
+    end
+
+    context 'when the custom role grants no conversation permissions' do
+      before do
+        custom_role = create(:custom_role, account: account, permissions: ['contact_manage'])
+        account_user.update!(custom_role: custom_role)
+      end
+
+      it 'does not create a notification' do
+        conversation = create(:conversation, account: account, inbox: inbox, assignee: agent)
+
+        expect { build_notification(conversation) }.not_to(change { agent.notifications.count })
+      end
+    end
+  end
+end
diff --git a/spec/enterprise/models/sla_event_spec.rb b/spec/enterprise/models/sla_event_spec.rb
index 8a839d626..0600fb80f 100644
--- a/spec/enterprise/models/sla_event_spec.rb
+++ b/spec/enterprise/models/sla_event_spec.rb
@@ -55,6 +55,7 @@ RSpec.describe SlaEvent, type: :model do
     before do
       # to ensure notifications are not sent to other users
       create(:user, account: account)
+      create(:inbox_member, inbox: inbox, user: assignee)
       create(:inbox_member, inbox: inbox, user: participant)
       create(:conversation_participant, conversation: conversation, user: participant)
     end
diff --git a/spec/services/messages/mention_service_spec.rb b/spec/services/messages/mention_service_spec.rb
index a7bddcc4e..e04a0e19b 100644
--- a/spec/services/messages/mention_service_spec.rb
+++ b/spec/services/messages/mention_service_spec.rb
@@ -117,6 +117,26 @@ describe Messages::MentionService do
 
         expect(conversation.conversation_participants.map(&:user_id)).to include(first_agent.id)
       end
+
+      it 'adds the mentioned user as a participant before generating the notification' do
+        message = build(
+          :message,
+          conversation: conversation,
+          account: account,
+          content: "hi (mention://user/#{first_agent.id}/#{first_agent.name})",
+          private: true
+        )
+
+        participant_user_ids_when_notified = nil
+        allow(NotificationBuilder).to receive(:new) do |**_kwargs|
+          participant_user_ids_when_notified = conversation.conversation_participants.reload.map(&:user_id)
+          builder
+        end
+
+        described_class.new(message: message).perform
+
+        expect(participant_user_ids_when_notified).to include(first_agent.id)
+      end
     end
 
     context 'when message contains multiple user mentions' do

From 58fdd206250276f5a9d03c067b6149868f6b3a32 Mon Sep 17 00:00:00 2001
From: Tanmay Deep Sharma <32020192+tds-1@users.noreply.github.com>
Date: Tue, 12 May 2026 12:40:13 +0530
Subject: [PATCH 19/20] test(voice): WhatsApp Cloud Calling specs [5] (#14357)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Backend test coverage for the WhatsApp Cloud Calling pipeline introduced
in #14356. Stacked on top of that PR so the controller and service under
test exist when CI runs.

## Closes
- Replaces #14348 (which was based on the abandoned \`feature/pla-150\`)

## What's covered

-
\`spec/enterprise/controllers/api/v1/accounts/whatsapp_calls_controller_spec.rb\`
(new, ~210 lines)
- \`show / accept / reject / terminate / initiate / upload_recording\`
happy paths
- 422 paths: missing sdp_offer, missing recording, calling_disabled
inbox, missing contact phone, ringing-state guards, AlreadyAccepted,
NotRinging, CallFailed
- 138006 (no permission) → throttled opt-in template send under
conversation lock; idempotency on retry
  - \`upload_recording\` idempotency guard (\`already_uploaded\`)

- \`spec/enterprise/services/whatsapp/call_service_spec.rb\` (new, ~135
lines)
- State machine: ringing → in_progress → completed; ringing → failed
(reject); ringing → no_answer (terminate)
- Lock contention: concurrent terminate during accept doesn't corrupt
the message/conversation broadcast
- Provider failure paths surface as \`Voice::CallErrors::CallFailed\`
(transport and business)

- \`spec/models/channel/whatsapp_spec.rb\` — extends existing file with
\`voice_enabled?\` matrix (provider × source × calling_enabled)

## Verification

- 77/77 examples pass locally on this branch (controller + service +
channel + incoming-call + permission-reply + open-ai message builder)
- RuboCop clean

## Stack

- Backend: #14356 (\`feat/whatsapp-call-meta-bridge\` — base of this PR)
- FE: #14346 (\`feat/whatsapp-call-ui\`)

---------

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---
 .../whatsapp_calls_controller_spec.rb         | 209 ++++++++++++++++++
 .../voice/inbound_call_builder_spec.rb        |   4 +-
 .../services/whatsapp/call_service_spec.rb    | 136 ++++++++++++
 spec/models/channel/whatsapp_spec.rb          |  39 ++++
 4 files changed, 387 insertions(+), 1 deletion(-)
 create mode 100644 spec/enterprise/controllers/api/v1/accounts/whatsapp_calls_controller_spec.rb
 create mode 100644 spec/enterprise/services/whatsapp/call_service_spec.rb

diff --git a/spec/enterprise/controllers/api/v1/accounts/whatsapp_calls_controller_spec.rb b/spec/enterprise/controllers/api/v1/accounts/whatsapp_calls_controller_spec.rb
new file mode 100644
index 000000000..264f428ce
--- /dev/null
+++ b/spec/enterprise/controllers/api/v1/accounts/whatsapp_calls_controller_spec.rb
@@ -0,0 +1,209 @@
+require 'rails_helper'
+
+RSpec.describe 'WhatsApp Calls API', type: :request do
+  let(:account) { create(:account) }
+  let(:agent) { create(:user, account: account, role: :agent) }
+  let(:channel) do
+    create(:channel_whatsapp, provider: 'whatsapp_cloud', account: account,
+                              validate_provider_config: false, sync_templates: false)
+  end
+  let(:inbox) { channel.inbox }
+  let(:conversation) { create(:conversation, account: account, inbox: inbox) }
+  let(:call) do
+    create(:call, account: account, inbox: inbox, conversation: conversation, contact: conversation.contact,
+                  provider: :whatsapp, direction: :incoming, status: 'ringing', provider_call_id: 'wacid_abc')
+  end
+  let(:provider_service) { instance_double(Whatsapp::Providers::WhatsappCloudService) }
+
+  before do
+    account.enable_features!('channel_voice')
+    channel.provider_config = channel.provider_config.merge('source' => 'embedded_signup', 'calling_enabled' => true)
+    channel.save!
+    create(:inbox_member, user: agent, inbox: inbox)
+    allow(Whatsapp::Providers::WhatsappCloudService).to receive(:new).and_return(provider_service)
+  end
+
+  describe 'GET /api/v1/accounts/:account_id/whatsapp_calls/:id' do
+    it 'returns the call payload' do
+      get "/api/v1/accounts/#{account.id}/whatsapp_calls/#{call.id}", headers: agent.create_new_auth_token
+
+      expect(response).to have_http_status(:ok)
+      body = response.parsed_body
+      expect(body['id']).to eq(call.id)
+      expect(body['call_id']).to eq('wacid_abc')
+      expect(body['provider']).to eq('whatsapp')
+    end
+
+    it 'returns 401 when unauthenticated' do
+      get "/api/v1/accounts/#{account.id}/whatsapp_calls/#{call.id}"
+      expect(response).to have_http_status(:unauthorized)
+    end
+  end
+
+  describe 'POST /api/v1/accounts/:account_id/whatsapp_calls/:id/accept' do
+    it 'forwards SDP and returns the updated call payload' do
+      allow(provider_service).to receive(:pre_accept_call).and_return(true)
+      allow(provider_service).to receive(:accept_call).and_return(true)
+
+      post "/api/v1/accounts/#{account.id}/whatsapp_calls/#{call.id}/accept",
+           params: { sdp_answer: 'sdp_answer' }, headers: agent.create_new_auth_token
+
+      expect(response).to have_http_status(:ok)
+      expect(call.reload.status).to eq('in_progress')
+    end
+
+    it 'returns 422 when sdp_answer is missing' do
+      post "/api/v1/accounts/#{account.id}/whatsapp_calls/#{call.id}/accept",
+           headers: agent.create_new_auth_token
+
+      expect(response).to have_http_status(:unprocessable_entity)
+    end
+  end
+
+  describe 'POST /api/v1/accounts/:account_id/whatsapp_calls/:id/reject' do
+    it 'rejects the call via Meta and returns its new status' do
+      allow(provider_service).to receive(:reject_call).and_return(true)
+
+      post "/api/v1/accounts/#{account.id}/whatsapp_calls/#{call.id}/reject",
+           headers: agent.create_new_auth_token
+
+      expect(response).to have_http_status(:ok)
+      expect(call.reload.status).to eq('failed')
+    end
+  end
+
+  describe 'POST /api/v1/accounts/:account_id/whatsapp_calls/:id/terminate' do
+    it 'terminates the call via Meta and returns its new status' do
+      call.update!(status: 'in_progress')
+      allow(provider_service).to receive(:terminate_call).and_return(true)
+
+      post "/api/v1/accounts/#{account.id}/whatsapp_calls/#{call.id}/terminate",
+           headers: agent.create_new_auth_token
+
+      expect(response).to have_http_status(:ok)
+      expect(call.reload.status).to eq('completed')
+    end
+  end
+
+  describe 'POST /api/v1/accounts/:account_id/whatsapp_calls/initiate' do
+    let(:contact) { create(:contact, account: account, phone_number: '+15551234567') }
+    let!(:contact_inbox) { create(:contact_inbox, contact: contact, inbox: inbox, source_id: '15551234567') }
+    let(:initiate_conversation) do
+      create(:conversation, account: account, inbox: inbox, contact: contact, contact_inbox: contact_inbox)
+    end
+
+    it 'creates an outbound Call and returns calling status' do
+      allow(provider_service).to receive(:initiate_call).and_return({ 'calls' => [{ 'id' => 'wacid_outbound' }] })
+
+      post "/api/v1/accounts/#{account.id}/whatsapp_calls/initiate",
+           params: { conversation_id: initiate_conversation.display_id, sdp_offer: 'sdp_offer' },
+           headers: agent.create_new_auth_token
+
+      expect(response).to have_http_status(:ok)
+      expect(response.parsed_body).to include('status' => 'calling', 'call_id' => 'wacid_outbound')
+      expect(Call.find_by(provider_call_id: 'wacid_outbound')).to have_attributes(direction: 'outgoing', status: 'ringing')
+    end
+
+    it 'sends a permission request and records the wamid when Meta returns NoCallPermission' do
+      allow(provider_service).to receive(:initiate_call).and_raise(Voice::CallErrors::NoCallPermission)
+      allow(provider_service).to receive(:send_call_permission_request).and_return({ 'messages' => [{ 'id' => 'wamid.req_xyz' }] })
+
+      post "/api/v1/accounts/#{account.id}/whatsapp_calls/initiate",
+           params: { conversation_id: initiate_conversation.display_id, sdp_offer: 'sdp_offer' },
+           headers: agent.create_new_auth_token
+
+      # Controller deliberately returns 422 so clients can't mistake the permission-template path for a successful dial.
+      expect(response).to have_http_status(:unprocessable_entity)
+      expect(response.parsed_body['status']).to eq('permission_requested')
+      attrs = initiate_conversation.reload.additional_attributes
+      expect(attrs['call_permission_requested_at']).to be_present
+      expect(attrs['call_permission_request_message_id']).to eq('wamid.req_xyz')
+    end
+
+    it 'returns permission_request_failed when send_call_permission_request raises a transport error' do
+      allow(provider_service).to receive(:initiate_call).and_raise(Voice::CallErrors::NoCallPermission)
+      allow(provider_service).to receive(:send_call_permission_request).and_raise(Faraday::TimeoutError)
+
+      post "/api/v1/accounts/#{account.id}/whatsapp_calls/initiate",
+           params: { conversation_id: initiate_conversation.display_id, sdp_offer: 'sdp_offer' },
+           headers: agent.create_new_auth_token
+
+      expect(response).to have_http_status(:unprocessable_entity)
+      expect(response.parsed_body['error']).to eq(I18n.t('errors.whatsapp.calls.permission_request_failed'))
+    end
+
+    it 'returns 422 when sdp_offer is missing' do
+      post "/api/v1/accounts/#{account.id}/whatsapp_calls/initiate",
+           params: { conversation_id: initiate_conversation.display_id },
+           headers: agent.create_new_auth_token
+
+      expect(response).to have_http_status(:unprocessable_entity)
+    end
+
+    it 'returns 422 when Meta raises CallFailed for non-permission errors' do
+      allow(provider_service).to receive(:initiate_call).and_raise(Voice::CallErrors::CallFailed, 'Meta error')
+
+      post "/api/v1/accounts/#{account.id}/whatsapp_calls/initiate",
+           params: { conversation_id: initiate_conversation.display_id, sdp_offer: 'sdp_offer' },
+           headers: agent.create_new_auth_token
+
+      expect(response).to have_http_status(:unprocessable_entity)
+      expect(response.parsed_body['error']).to eq('Meta error')
+    end
+
+    it 'returns 422 when the conversation contact has no phone number' do
+      contact.update!(phone_number: nil)
+
+      post "/api/v1/accounts/#{account.id}/whatsapp_calls/initiate",
+           params: { conversation_id: initiate_conversation.display_id, sdp_offer: 'sdp_offer' },
+           headers: agent.create_new_auth_token
+
+      expect(response).to have_http_status(:unprocessable_entity)
+      expect(response.parsed_body['error']).to eq(I18n.t('errors.whatsapp.calls.contact_phone_required'))
+    end
+
+    it 'returns 422 when the conversation belongs to a non-WhatsApp inbox' do
+      twilio_channel = create(:channel_twilio_sms, :with_voice, account: account, phone_number: '+15551239998')
+      create(:inbox_member, user: agent, inbox: twilio_channel.inbox)
+      twilio_conversation = create(:conversation, account: account, inbox: twilio_channel.inbox)
+
+      post "/api/v1/accounts/#{account.id}/whatsapp_calls/initiate",
+           params: { conversation_id: twilio_conversation.display_id, sdp_offer: 'sdp_offer' },
+           headers: agent.create_new_auth_token
+
+      expect(response).to have_http_status(:unprocessable_entity)
+      expect(response.parsed_body['error']).to eq(I18n.t('errors.whatsapp.calls.not_enabled'))
+    end
+  end
+
+  describe 'POST /api/v1/accounts/:account_id/whatsapp_calls/:id/upload_recording' do
+    before do
+      message = create(:message, conversation: conversation, account: account, inbox: inbox,
+                                 content_type: 'voice_call', message_type: 'incoming')
+      call.update!(message_id: message.id)
+    end
+
+    it 'attaches the recording to the call message' do
+      file = fixture_file_upload(Rails.root.join('spec/assets/sample.mp3'), 'audio/mpeg')
+
+      expect do
+        post "/api/v1/accounts/#{account.id}/whatsapp_calls/#{call.id}/upload_recording",
+             params: { recording: file }, headers: agent.create_new_auth_token
+      end.to change { call.message.attachments.count }.by(1)
+
+      expect(response).to have_http_status(:ok)
+      expect(response.parsed_body['status']).to eq('uploaded')
+    end
+
+    it 'is idempotent: returns already_uploaded if an audio attachment exists' do
+      call.message.attachments.create!(account_id: account.id, file_type: :audio,
+                                       file: fixture_file_upload(Rails.root.join('spec/assets/sample.mp3'), 'audio/mpeg'))
+
+      post "/api/v1/accounts/#{account.id}/whatsapp_calls/#{call.id}/upload_recording",
+           params: { recording: fixture_file_upload(Rails.root.join('spec/assets/sample.mp3'), 'audio/mpeg') },
+           headers: agent.create_new_auth_token
+
+      expect(response.parsed_body['status']).to eq('already_uploaded')
+    end
+  end
+end
diff --git a/spec/enterprise/services/voice/inbound_call_builder_spec.rb b/spec/enterprise/services/voice/inbound_call_builder_spec.rb
index 3065875f2..e36c9a1b7 100644
--- a/spec/enterprise/services/voice/inbound_call_builder_spec.rb
+++ b/spec/enterprise/services/voice/inbound_call_builder_spec.rb
@@ -103,7 +103,7 @@ RSpec.describe Voice::InboundCallBuilder do
   context 'when the WhatsApp wa_id needs Brazil normalization to match an existing ContactInbox' do
     let(:whatsapp_channel) do
       create(:channel_whatsapp, account: account, provider: 'whatsapp_cloud',
-                                provider_config: { 'phone_number_id' => '123', 'calling_enabled' => true },
+                                provider_config: { 'phone_number_id' => '123', 'source' => 'embedded_signup', 'calling_enabled' => true },
                                 validate_provider_config: false, sync_templates: false)
     end
     let(:whatsapp_inbox) { whatsapp_channel.inbox }
@@ -112,6 +112,8 @@ RSpec.describe Voice::InboundCallBuilder do
       create(:contact_inbox, contact: stored_contact, inbox: whatsapp_inbox, source_id: '5541988887777')
     end
 
+    before { account.enable_features!('channel_voice') }
+
     it 'reuses the contact via normalized wa_id rather than forking a new ContactInbox' do
       call = described_class.perform!(
         inbox: whatsapp_inbox,
diff --git a/spec/enterprise/services/whatsapp/call_service_spec.rb b/spec/enterprise/services/whatsapp/call_service_spec.rb
new file mode 100644
index 000000000..a17f7572a
--- /dev/null
+++ b/spec/enterprise/services/whatsapp/call_service_spec.rb
@@ -0,0 +1,136 @@
+require 'rails_helper'
+
+describe Whatsapp::CallService do
+  let(:account) { create(:account) }
+  let(:channel) do
+    create(:channel_whatsapp, provider: 'whatsapp_cloud', account: account,
+                              validate_provider_config: false, sync_templates: false)
+  end
+  let(:inbox) { channel.inbox }
+  let(:agent) { create(:user, account: account) }
+  let(:conversation) { create(:conversation, account: account, inbox: inbox) }
+  let(:call) do
+    create(:call, account: account, inbox: inbox, conversation: conversation, contact: conversation.contact,
+                  provider: :whatsapp, direction: :incoming, status: 'ringing', provider_call_id: 'wacid_abc')
+  end
+  let(:provider_service) { instance_double(Whatsapp::Providers::WhatsappCloudService) }
+
+  before do
+    channel.provider_config = channel.provider_config.merge('calling_enabled' => true)
+    channel.save!
+    allow(channel).to receive(:provider_service).and_return(provider_service)
+    allow(inbox).to receive(:channel).and_return(channel)
+    allow(call).to receive(:inbox).and_return(inbox)
+    allow(ActionCable.server).to receive(:broadcast)
+  end
+
+  describe '#accept' do
+    let(:sdp_answer) { "v=0\r\n...sdp..." }
+
+    before do
+      allow(provider_service).to receive(:pre_accept_call).and_return(true)
+      allow(provider_service).to receive(:accept_call).and_return(true)
+    end
+
+    it 'forwards the SDP answer to Meta and transitions the call to in_progress' do
+      described_class.new(call: call, agent: agent, sdp_answer: sdp_answer).accept
+
+      expect(provider_service).to have_received(:pre_accept_call).with('wacid_abc', sdp_answer)
+      expect(provider_service).to have_received(:accept_call).with('wacid_abc', sdp_answer)
+      expect(call.reload).to have_attributes(status: 'in_progress', accepted_by_agent_id: agent.id, started_at: be_present)
+      expect(call.meta['sdp_answer']).to eq(sdp_answer)
+      expect(ActionCable.server).to have_received(:broadcast).with(
+        "account_#{account.id}", hash_including(event: 'voice_call.accepted')
+      )
+    end
+
+    it 'claims the conversation when no assignee is set' do
+      described_class.new(call: call, agent: agent, sdp_answer: sdp_answer).accept
+
+      expect(conversation.reload.assignee_id).to eq(agent.id)
+    end
+
+    it 'raises AlreadyAccepted when another agent has already accepted the call' do
+      call.update!(status: 'in_progress')
+
+      expect { described_class.new(call: call, agent: agent, sdp_answer: sdp_answer).accept }
+        .to raise_error(StandardError) { |error| expect(error.class.name).to eq('Voice::CallErrors::AlreadyAccepted') }
+    end
+
+    it 'raises NotRinging when the call has reached a terminal state' do
+      call.update!(status: 'completed')
+
+      expect { described_class.new(call: call, agent: agent, sdp_answer: sdp_answer).accept }
+        .to raise_error(StandardError) { |error| expect(error.class.name).to eq('Voice::CallErrors::NotRinging') }
+    end
+
+    it 'raises CallFailed when sdp_answer is missing' do
+      expect { described_class.new(call: call, agent: agent, sdp_answer: nil).accept }
+        .to raise_error(StandardError) do |error|
+          expect(error.class.name).to eq('Voice::CallErrors::CallFailed')
+          expect(error.message).to eq('sdp_answer is required')
+        end
+    end
+
+    it 'wraps Meta transport exceptions as CallFailed and leaves the call ringing' do
+      allow(provider_service).to receive(:pre_accept_call).and_raise(Faraday::TimeoutError)
+
+      expect { described_class.new(call: call, agent: agent, sdp_answer: sdp_answer).accept }
+        .to raise_error(StandardError) { |error| expect(error.class.name).to eq('Voice::CallErrors::CallFailed') }
+      expect(call.reload.status).to eq('ringing')
+    end
+  end
+
+  describe '#reject' do
+    before { allow(provider_service).to receive(:reject_call).and_return(true) }
+
+    it 'tells Meta to reject and finalizes the call as failed' do
+      described_class.new(call: call, agent: agent).reject
+
+      expect(provider_service).to have_received(:reject_call).with('wacid_abc')
+      expect(call.reload.status).to eq('failed')
+      expect(ActionCable.server).to have_received(:broadcast).with(
+        "account_#{account.id}", hash_including(event: 'voice_call.ended', data: hash_including(status: 'failed'))
+      )
+    end
+
+    it 'is a no-op for already-terminal calls' do
+      call.update!(status: 'completed')
+
+      described_class.new(call: call, agent: agent).reject
+
+      expect(provider_service).not_to have_received(:reject_call)
+    end
+
+    it 'raises CallFailed and leaves the call ringing when Meta rejects the request' do
+      allow(provider_service).to receive(:reject_call).and_return(false)
+
+      expect { described_class.new(call: call, agent: agent).reject }
+        .to raise_error(StandardError) { |error| expect(error.class.name).to eq('Voice::CallErrors::CallFailed') }
+      expect(call.reload.status).to eq('ringing')
+    end
+  end
+
+  describe '#terminate' do
+    before { allow(provider_service).to receive(:terminate_call).and_return(true) }
+
+    it 'finalizes an in-progress call as completed' do
+      call.update!(status: 'in_progress')
+
+      described_class.new(call: call, agent: agent).terminate
+
+      expect(provider_service).to have_received(:terminate_call).with('wacid_abc')
+      expect(call.reload.status).to eq('completed')
+      expect(call.meta['ended_at']).to be_present
+      expect(ActionCable.server).to have_received(:broadcast).with(
+        "account_#{account.id}", hash_including(event: 'voice_call.ended')
+      )
+    end
+
+    it 'finalizes a still-ringing call as no_answer when the agent hangs up before the contact picks up' do
+      described_class.new(call: call, agent: agent).terminate
+
+      expect(call.reload.status).to eq('no_answer')
+    end
+  end
+end
diff --git a/spec/models/channel/whatsapp_spec.rb b/spec/models/channel/whatsapp_spec.rb
index dcc010d88..9afc21f7c 100644
--- a/spec/models/channel/whatsapp_spec.rb
+++ b/spec/models/channel/whatsapp_spec.rb
@@ -209,4 +209,43 @@ RSpec.describe Channel::Whatsapp do
       end
     end
   end
+
+  describe '#voice_enabled?' do
+    let(:account) { create(:account) }
+
+    before { account.enable_features!('channel_voice') }
+
+    it 'returns true for embedded-signup whatsapp_cloud channels with calling_enabled' do
+      channel = create(:channel_whatsapp, account: account, provider: 'whatsapp_cloud',
+                                          validate_provider_config: false, sync_templates: false)
+      channel.update!(provider_config: channel.provider_config.merge('source' => 'embedded_signup', 'calling_enabled' => true))
+
+      expect(channel.voice_enabled?).to be true
+    end
+
+    it 'returns false for whatsapp_cloud channels without embedded_signup source' do
+      channel = create(:channel_whatsapp, account: account, provider: 'whatsapp_cloud',
+                                          validate_provider_config: false, sync_templates: false)
+      channel.update!(provider_config: channel.provider_config.merge('source' => 'manual', 'calling_enabled' => true))
+
+      expect(channel.voice_enabled?).to be false
+    end
+
+    it 'returns false for default-provider channels (360dialog) even with calling_enabled' do
+      channel = create(:channel_whatsapp, account: account, provider: 'default',
+                                          validate_provider_config: false, sync_templates: false)
+      channel.update!(provider_config: channel.provider_config.merge('source' => 'embedded_signup', 'calling_enabled' => true))
+
+      expect(channel.voice_enabled?).to be false
+    end
+
+    it 'returns false when the channel_voice feature is disabled on the account' do
+      account.disable_features!('channel_voice')
+      channel = create(:channel_whatsapp, account: account, provider: 'whatsapp_cloud',
+                                          validate_provider_config: false, sync_templates: false)
+      channel.update!(provider_config: channel.provider_config.merge('source' => 'embedded_signup', 'calling_enabled' => true))
+
+      expect(channel.voice_enabled?).to be false
+    end
+  end
 end

From 71cc5168beef55f0fae07e3f0d7f9dd7dea1692d Mon Sep 17 00:00:00 2001
From: Muhsin Keloth <muhsinkeramam@gmail.com>
Date: Tue, 12 May 2026 13:03:40 +0400
Subject: [PATCH 20/20] feat(linear): Auto link Linear issues from private
 notes (#14405)
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

When an agent pastes a Linear issue URL into a private note on a
conversation, Chatwoot now links the issue to the conversation
automatically — no need to click "Link to Linear issue" first. The
standard activity message ("X linked Linear issue ABC-123") is posted
just like a manual link.

Fixes
[CW-7032](https://linear.app/chatwoot/issue/CW-7032/if-someone-post-a-linear-url-in-the-private-notes-automatically-link)

---------

Co-authored-by: Muhsin <12408980+muhsin-k@users.noreply.github.com>
Co-authored-by: Sojan Jose <sojan@pepalo.com>
---
 app/jobs/hook_job.rb                          |  27 ++-
 app/listeners/hook_listener.rb                |   3 +-
 config/locales/en.yml                         |   1 +
 lib/integrations/linear/auto_link_service.rb  |  93 +++++++++
 lib/linear/queries.rb                         |   1 +
 spec/jobs/hook_job_spec.rb                    |   7 +
 .../linear/auto_link_service_spec.rb          | 178 ++++++++++++++++++
 spec/listeners/hook_listener_spec.rb          |   7 +
 8 files changed, 306 insertions(+), 11 deletions(-)
 create mode 100644 lib/integrations/linear/auto_link_service.rb
 create mode 100644 spec/lib/integrations/linear/auto_link_service_spec.rb

diff --git a/app/jobs/hook_job.rb b/app/jobs/hook_job.rb
index deb7c81d8..eff80844f 100644
--- a/app/jobs/hook_job.rb
+++ b/app/jobs/hook_job.rb
@@ -3,19 +3,19 @@ class HookJob < MutexApplicationJob
 
   queue_as :medium
 
+  INTEGRATION_PROCESSORS = {
+    'slack' => :process_slack_integration,
+    'dialogflow' => :process_dialogflow_integration,
+    'google_translate' => :google_translate_integration,
+    'leadsquared' => :process_leadsquared_integration_with_lock,
+    'linear' => :process_linear_integration
+  }.freeze
+
   def perform(hook, event_name, event_data = {})
     return if hook.disabled?
 
-    case hook.app_id
-    when 'slack'
-      process_slack_integration(hook, event_name, event_data)
-    when 'dialogflow'
-      process_dialogflow_integration(hook, event_name, event_data)
-    when 'google_translate'
-      google_translate_integration(hook, event_name, event_data)
-    when 'leadsquared'
-      process_leadsquared_integration_with_lock(hook, event_name, event_data)
-    end
+    processor = INTEGRATION_PROCESSORS[hook.app_id]
+    send(processor, hook, event_name, event_data) if processor
   rescue StandardError => e
     Rails.logger.error e
   end
@@ -57,6 +57,13 @@ class HookJob < MutexApplicationJob
     Integrations::GoogleTranslate::DetectLanguageService.new(hook: hook, message: message).perform
   end
 
+  def process_linear_integration(hook, event_name, event_data)
+    return unless event_name == 'message.created'
+
+    message = event_data[:message]
+    Integrations::Linear::AutoLinkService.new(account: hook.account, message: message).perform
+  end
+
   def process_leadsquared_integration_with_lock(hook, event_name, event_data)
     # Why do we need a mutex here? glad you asked
     # When a new conversation is created. We get a contact created event, immediately followed by
diff --git a/app/listeners/hook_listener.rb b/app/listeners/hook_listener.rb
index 6176d53dd..7c55b76d1 100644
--- a/app/listeners/hook_listener.rb
+++ b/app/listeners/hook_listener.rb
@@ -62,7 +62,8 @@ class HookListener < BaseListener
       'slack' => ['message.created', 'message.updated'],
       'dialogflow' => ['message.created', 'message.updated'],
       'google_translate' => ['message.created'],
-      'leadsquared' => ['contact.updated', 'conversation.created', 'conversation.resolved']
+      'leadsquared' => ['contact.updated', 'conversation.created', 'conversation.resolved'],
+      'linear' => ['message.created']
     }
 
     return false unless supported_events_map.key?(hook.app_id)
diff --git a/config/locales/en.yml b/config/locales/en.yml
index f7b98a47b..c590f8be9 100644
--- a/config/locales/en.yml
+++ b/config/locales/en.yml
@@ -367,6 +367,7 @@ en:
       name: 'Linear'
       short_description: 'Create and link Linear issues directly from conversations.'
       description: 'Create issues in Linear directly from your conversation window. Alternatively, link existing Linear issues for a more streamlined and efficient issue tracking process.'
+      attachment_link_title: 'Conversation (#%{conversation_id}) with %{name}'
     notion:
       name: 'Notion'
       short_description: 'Integrate databases, documents and pages directly with Captain.'
diff --git a/lib/integrations/linear/auto_link_service.rb b/lib/integrations/linear/auto_link_service.rb
new file mode 100644
index 000000000..a980e56e4
--- /dev/null
+++ b/lib/integrations/linear/auto_link_service.rb
@@ -0,0 +1,93 @@
+class Integrations::Linear::AutoLinkService
+  pattr_initialize [:account!, :message!]
+
+  LINEAR_URL_REGEX = %r{https?://linear\.app/[^/\s]+/issue/[A-Z][A-Z0-9_]+-\d+(?:/[^\s)]*)?}
+  IDENTIFIER_REGEX = %r{/issue/([A-Z][A-Z0-9_]+-\d+)}i
+  WORKSPACE_REGEX = %r{//linear\.app/([^/\s]+)/}i
+
+  def perform
+    return unless valid_message?
+
+    attempt_link
+  end
+
+  private
+
+  def valid_message?
+    message.private? && message.content.present? && message.sender.is_a?(User)
+  end
+
+  def attempt_link
+    linear_url = message.content[LINEAR_URL_REGEX]
+    return if linear_url.blank?
+
+    identifier = linear_url[IDENTIFIER_REGEX, 1]&.upcase
+    workspace = linear_url[WORKSPACE_REGEX, 1]&.downcase
+    return if identifier.blank? || workspace.blank? || already_linked?(identifier)
+
+    finalize_link(workspace, identifier)
+  end
+
+  def finalize_link(workspace, identifier)
+    node_id = resolve_node_id(workspace, identifier)
+    return if node_id.blank?
+    return unless link_to_linear(node_id, identifier)
+
+    post_activity_message(identifier)
+  end
+
+  def already_linked?(identifier)
+    response = processor.linked_issues(conversation_link)
+    return false if response[:error]
+
+    response[:data].any? { |attachment| attachment.dig('issue', 'identifier') == identifier }
+  end
+
+  def resolve_node_id(workspace, identifier)
+    response = processor.search_issue(identifier)
+    return if response[:error]
+
+    node = response[:data].find do |issue|
+      issue['identifier'] == identifier && node_workspace(issue) == workspace
+    end
+    node && node['id']
+  end
+
+  def node_workspace(node)
+    node['url']&.match(WORKSPACE_REGEX)&.[](1)&.downcase
+  end
+
+  def link_to_linear(node_id, identifier)
+    response = processor.link_issue(conversation_link, node_id, attachment_title, message.sender)
+    if response[:error].present?
+      Rails.logger.warn("[Linear::AutoLinkService] link_issue failed for #{identifier}: #{response[:error]}")
+      return false
+    end
+    true
+  end
+
+  def attachment_title
+    I18n.t(
+      'integration_apps.linear.attachment_link_title',
+      conversation_id: message.conversation.display_id,
+      name: message.conversation.contact&.name
+    )
+  end
+
+  def post_activity_message(identifier)
+    Linear::ActivityMessageService.new(
+      conversation: message.conversation,
+      action_type: :issue_linked,
+      user: message.sender,
+      issue_data: { id: identifier }
+    ).perform
+  end
+
+  def conversation_link
+    "#{ENV.fetch('FRONTEND_URL', nil)}/app/accounts/#{message.account_id}/conversations/#{message.conversation.display_id}"
+  end
+
+  def processor
+    @processor ||= Integrations::Linear::ProcessorService.new(account: account)
+  end
+end
diff --git a/lib/linear/queries.rb b/lib/linear/queries.rb
index cc705625e..7d7c9ae97 100644
--- a/lib/linear/queries.rb
+++ b/lib/linear/queries.rb
@@ -54,6 +54,7 @@ module Linear::Queries
             title
             description
             identifier
+            url
             state {
               name
               color
diff --git a/spec/jobs/hook_job_spec.rb b/spec/jobs/hook_job_spec.rb
index b55b05838..c6faebd86 100644
--- a/spec/jobs/hook_job_spec.rb
+++ b/spec/jobs/hook_job_spec.rb
@@ -65,6 +65,13 @@ RSpec.describe HookJob do
       expect(Integrations::GoogleTranslate::DetectLanguageService).to receive(:new).with(hook: hook, message: event_data[:message])
       described_class.perform_now(hook, event_name, event_data)
     end
+
+    it "calls Integrations::Linear::AutoLinkService when it's a linear hook" do
+      hook = create(:integrations_hook, :linear, account: account)
+      allow(Integrations::Linear::AutoLinkService).to receive(:new).and_return(process_service)
+      expect(Integrations::Linear::AutoLinkService).to receive(:new).with(account: account, message: event_data[:message])
+      described_class.perform_now(hook, event_name, event_data)
+    end
   end
 
   context 'when handleable events like message.updated for slack' do
diff --git a/spec/lib/integrations/linear/auto_link_service_spec.rb b/spec/lib/integrations/linear/auto_link_service_spec.rb
new file mode 100644
index 000000000..f90c5f060
--- /dev/null
+++ b/spec/lib/integrations/linear/auto_link_service_spec.rb
@@ -0,0 +1,178 @@
+require 'rails_helper'
+
+describe Integrations::Linear::AutoLinkService do
+  let(:account) { create(:account) }
+  let(:user) { create(:user, account: account) }
+  let(:inbox) { create(:inbox, account: account) }
+  let(:conversation) { create(:conversation, account: account, inbox: inbox) }
+  let(:processor) { instance_double(Integrations::Linear::ProcessorService) }
+  let(:activity_service) { instance_double(Linear::ActivityMessageService, perform: true) }
+
+  let(:linear_url) { 'https://linear.app/chatwoot/issue/CW-1234/some-slug' }
+  let(:identifier) { 'CW-1234' }
+  let(:node_id) { 'linear-node-id-1' }
+  let(:search_response) do
+    { data: [{ 'id' => node_id, 'identifier' => identifier, 'title' => 'Issue title',
+               'url' => 'https://linear.app/chatwoot/issue/CW-1234/issue-title' }] }
+  end
+
+  before do
+    allow(Integrations::Linear::ProcessorService).to receive(:new).with(account: account).and_return(processor)
+    allow(Linear::ActivityMessageService).to receive(:new).and_return(activity_service)
+    allow(processor).to receive(:linked_issues).and_return({ data: [] })
+    allow(processor).to receive(:search_issue).and_return(search_response)
+    allow(processor).to receive(:link_issue).and_return({ data: { id: node_id, link_id: 'attachment-1' } })
+  end
+
+  def build_private_note(content)
+    create(:message,
+           account: account,
+           inbox: inbox,
+           conversation: conversation,
+           sender: user,
+           message_type: :outgoing,
+           private: true,
+           content: content)
+  end
+
+  describe '#perform' do
+    context 'when the message is not a private note' do
+      it 'does no work' do
+        message = create(:message, account: account, inbox: inbox, conversation: conversation,
+                                   sender: user, message_type: :outgoing, private: false,
+                                   content: "see #{linear_url}")
+
+        described_class.new(account: account, message: message).perform
+
+        expect(processor).not_to have_received(:linked_issues)
+        expect(processor).not_to have_received(:search_issue)
+        expect(processor).not_to have_received(:link_issue)
+        expect(Linear::ActivityMessageService).not_to have_received(:new)
+      end
+    end
+
+    context 'when the sender is not a User' do
+      it 'does no work' do
+        contact = create(:contact, account: account)
+        message = create(:message, account: account, inbox: inbox, conversation: conversation,
+                                   sender: contact, message_type: :incoming, private: true,
+                                   content: "see #{linear_url}")
+
+        described_class.new(account: account, message: message).perform
+
+        expect(processor).not_to have_received(:link_issue)
+        expect(Linear::ActivityMessageService).not_to have_received(:new)
+      end
+    end
+
+    context 'when the private note has no Linear URL' do
+      it 'does no work' do
+        message = build_private_note('just a regular note with no link')
+
+        described_class.new(account: account, message: message).perform
+
+        expect(processor).not_to have_received(:link_issue)
+        expect(Linear::ActivityMessageService).not_to have_received(:new)
+      end
+    end
+
+    context 'when the issue identifier is already linked from this conversation' do
+      it 'skips silently' do
+        message = build_private_note("see #{linear_url}")
+        allow(processor).to receive(:linked_issues).and_return(
+          { data: [{ 'id' => 'attachment-prev', 'issue' => { 'id' => node_id, 'identifier' => identifier } }] }
+        )
+
+        described_class.new(account: account, message: message).perform
+
+        expect(processor).not_to have_received(:search_issue)
+        expect(processor).not_to have_received(:link_issue)
+        expect(Linear::ActivityMessageService).not_to have_received(:new)
+      end
+    end
+
+    context 'when Linear search returns no exact match for the identifier' do
+      it 'does not link' do
+        message = build_private_note("see #{linear_url}")
+        allow(processor).to receive(:search_issue).with(identifier).and_return(
+          { data: [{ 'id' => 'other', 'identifier' => 'OTHER-1', 'url' => 'https://linear.app/chatwoot/issue/OTHER-1' }] }
+        )
+
+        described_class.new(account: account, message: message).perform
+
+        expect(processor).not_to have_received(:link_issue)
+        expect(Linear::ActivityMessageService).not_to have_received(:new)
+      end
+    end
+
+    context 'when the matching issue belongs to a different Linear workspace' do
+      it 'does not link' do
+        message = build_private_note("see #{linear_url}")
+        allow(processor).to receive(:search_issue).with(identifier).and_return(
+          { data: [{ 'id' => node_id, 'identifier' => identifier,
+                     'url' => 'https://linear.app/other-workspace/issue/CW-1234' }] }
+        )
+
+        described_class.new(account: account, message: message).perform
+
+        expect(processor).not_to have_received(:link_issue)
+        expect(Linear::ActivityMessageService).not_to have_received(:new)
+      end
+    end
+
+    context 'when Linear search returns an error' do
+      it 'does not link' do
+        message = build_private_note("see #{linear_url}")
+        allow(processor).to receive(:search_issue).with(identifier).and_return({ error: 'boom' })
+
+        described_class.new(account: account, message: message).perform
+
+        expect(processor).not_to have_received(:link_issue)
+        expect(Linear::ActivityMessageService).not_to have_received(:new)
+      end
+    end
+
+    context 'when link_issue returns an error' do
+      it 'does not post the activity message' do
+        message = build_private_note("see #{linear_url}")
+        allow(processor).to receive(:link_issue).and_return({ error: 'nope' })
+
+        described_class.new(account: account, message: message).perform
+
+        expect(Linear::ActivityMessageService).not_to have_received(:new)
+      end
+    end
+
+    context 'when the private note contains a Linear URL' do
+      it 'links the issue and posts an activity message' do
+        message = build_private_note("Found it: #{linear_url}")
+
+        described_class.new(account: account, message: message).perform
+
+        expect(processor).to have_received(:link_issue).with(
+          a_string_matching(%r{/conversations/#{conversation.display_id}\z}),
+          node_id,
+          anything,
+          user
+        )
+        expect(Linear::ActivityMessageService).to have_received(:new).with(
+          conversation: conversation,
+          action_type: :issue_linked,
+          user: user,
+          issue_data: { id: identifier }
+        )
+        expect(activity_service).to have_received(:perform)
+      end
+
+      it 'links only the first Linear URL when multiple are present' do
+        second_url = 'https://linear.app/chatwoot/issue/CW-9999'
+        message = build_private_note("see #{linear_url} and #{second_url}")
+
+        described_class.new(account: account, message: message).perform
+
+        expect(processor).to have_received(:search_issue).with(identifier).once
+        expect(processor).not_to have_received(:search_issue).with('CW-9999')
+      end
+    end
+  end
+end
diff --git a/spec/listeners/hook_listener_spec.rb b/spec/listeners/hook_listener_spec.rb
index c048761d1..0f96ae4af 100644
--- a/spec/listeners/hook_listener_spec.rb
+++ b/spec/listeners/hook_listener_spec.rb
@@ -84,6 +84,13 @@ describe HookListener do
 
         listener.message_created(event)
       end
+
+      it 'enqueues the job for linear' do
+        hook = create(:integrations_hook, :linear, account: account)
+        expect(HookJob).to receive(:perform_later).with(hook, event_name, message: message, previous_changes: nil)
+
+        listener.message_created(event)
+      end
     end
 
     context 'with disabled hook' do