@ai-sdk/gateway 4.0.100 → 4.0.102

package/docs/00-ai-gateway.mdx

@@ -625,16 +625,20 @@ const result = await generateText({
 });
 
 // Get the generation ID from provider metadata
-const generationId = result.providerMetadata?.gateway?.generationId;
+const generationId = result.providerMetadata?.gateway?.generationId as
+  | string
+  | undefined;
 
 // Look up detailed generation info
-const generation = await gateway.getGenerationInfo({ id: generationId });
+if (generationId) {
+  const generation = await gateway.getGenerationInfo({ id: generationId });
 
-console.log(`Model: ${generation.model}`);
-console.log(`Cost: $${generation.totalCost.toFixed(6)}`);
-console.log(`Latency: ${generation.latency}ms`);
-console.log(`Prompt tokens: ${generation.promptTokens}`);
-console.log(`Completion tokens: ${generation.completionTokens}`);
+  console.log(`Model: ${generation.model}`);
+  console.log(`Cost: $${generation.totalCost.toFixed(6)}`);
+  console.log(`Latency: ${generation.latency}ms`);
+  console.log(`Prompt tokens: ${generation.promptTokens}`);
+  console.log(`Completion tokens: ${generation.completionTokens}`);
+}
 ```
 
 With `streamText`, you can capture the generation ID from the first chunk via `stream`:
@@ -1600,15 +1604,15 @@ const { text } = await generateText({
   prompt: 'Write a TypeScript haiku',
   providerOptions: {
     gateway: {
-      models: ['openai/gpt-5.4-nano', 'gemini-3.8-flash'], // Fallback models
+      models: ['openai/gpt-5.4-nano', 'google/gemini-3.8-flash'], // Fallback models
     } satisfies GatewayProviderOptions,
   },
 });
 
 // This will:
-// 1. Try openai/gpt-5.4 first
+// 1. Try openai/gpt-6-astra first
 // 2. If it fails, try openai/gpt-5.4-nano
-// 3. If that fails, try gemini-3-flash-preview
+// 3. If that fails, try google/gemini-3.8-flash
 // 4. Return the result from the first model that succeeds
 ```
 

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@ai-sdk/gateway",
   "private": false,
-  "version": "4.0.100",
+  "version": "4.0.102",
   "type": "module",
   "license": "Apache-2.0",
   "sideEffects": false,
@@ -30,15 +30,15 @@
     }
   },
   "dependencies": {
-    "@ai-sdk/provider": "4.0.19",
-    "@ai-sdk/provider-utils": "5.0.51",
+    "@ai-sdk/provider": "4.0.21",
+    "@ai-sdk/provider-utils": "5.0.53",
     "@vercel/oidc": "3.2.0"
   },
   "devDependencies": {
-    "@ai-sdk/test-server": "2.0.2",
+    "@ai-sdk/test-server": "2.0.3",
     "@types/node": "22.19.19",
     "@vercel/ai-tsconfig": "0.0.0",
-    "tsup": "^8.5.1",
+    "tsdown": "^0.23.0",
     "tsx": "4.23.12",
     "typescript": "5.8.3",
     "zod": "3.25.76"
@@ -66,9 +66,9 @@
     "ai"
   ],
   "scripts": {
-    "build": "pnpm clean && tsup --tsconfig tsconfig.build.json",
-    "build:watch": "pnpm clean && tsup --watch",
-    "clean": "del-cli dist docs *.tsbuildinfo",
+    "build": "pnpm clean && tsdown",
+    "build:watch": "pnpm clean && tsdown --watch",
+    "clean": "del-cli docs *.tsbuildinfo",
     "generate-model-settings": "tsx scripts/generate-model-settings.ts",
     "type-check": "tsc --build",
     "test": "pnpm test:node && pnpm test:edge",

package/src/gateway-evaluation-model-settings.ts

@@ -1 +1,4 @@
-export type GatewayEvaluationModelId = 'typesafe-ai/jev' | (string & {});
+export type GatewayEvaluationModelId =
+  | 'liquid/d1'
+  | 'typesafe-ai/jev'
+  | (string & {});

package/src/gateway-fetch-metadata.ts

@@ -10,6 +10,7 @@ import {
 import { z } from './zod';
 import { asGatewayError } from './errors';
 import type { GatewayConfig } from './gateway-config';
+import { VERCEL_AI_GATEWAY_TEAM_HEADER } from './gateway-headers';
 import {
   KNOWN_MODEL_TYPES,
   type GatewayLanguageModelEntry,
@@ -58,13 +59,27 @@ export class GatewayFetchMetadata {
   async getCredits(): Promise<GatewayCreditsResponse> {
     try {
       const baseUrl = new URL(this.config.baseURL);
+      const headers = this.config.headers
+        ? await resolve(this.config.headers)
+        : undefined;
+
+      // The credits endpoint selects the team from the `teamId` / `slug`
+      // query parameter rather than the team header, so forward it there.
+      // Without it, tokens that can access multiple teams (e.g. Vercel access
+      // tokens) cannot be scoped to a team and the request is rejected.
+      const url = new URL('/v1/credits', baseUrl.origin);
+      const teamIdOrSlug = getTeamIdOrSlug(headers);
+      if (teamIdOrSlug) {
+        url.searchParams.set(
+          teamIdOrSlug.startsWith('team_') ? 'teamId' : 'slug',
+          teamIdOrSlug,
+        );
+      }
 
       const { value } = await getFromApi({
-        url: `${baseUrl.origin}/v1/credits`,
+        url: url.toString(),
         validateUrl: false,
-        headers: this.config.headers
-          ? await resolve(this.config.headers)
-          : undefined,
+        headers,
         successfulResponseHandler: createJsonResponseHandler(
           gatewayCreditsResponseSchema,
         ),
@@ -82,6 +97,18 @@ export class GatewayFetchMetadata {
   }
 }
 
+function getTeamIdOrSlug(
+  headers: Record<string, string | undefined> | undefined,
+): string | undefined {
+  if (!headers) return undefined;
+  for (const [name, value] of Object.entries(headers)) {
+    if (name.toLowerCase() === VERCEL_AI_GATEWAY_TEAM_HEADER) {
+      return value?.trim() || undefined;
+    }
+  }
+  return undefined;
+}
+
 const gatewayAvailableModelsResponseSchema = lazySchema(() =>
   zodSchema(
     z.object({

package/src/gateway-language-model-settings.ts

@@ -98,6 +98,8 @@ export type GatewayModelId =
   | 'inclusionai/ling-3.0-flash-sante'
   | 'inclusionai/ling-3.0-flash-sante-free'
   | 'inclusionai/ling-3.0-flash-vl'
+  | 'inclusionai/ling-3.1-flash'
+  | 'inclusionai/ling-3.1-flash-free'
   | 'inference-net/schematron-v2-small'
   | 'inference-net/schematron-v2-turbo'
   | 'interfaze/interfaze-beta'
@@ -197,6 +199,8 @@ export type GatewayModelId =
   | 'openai/gpt-6-luna-fast'
   | 'openai/gpt-6-sol'
   | 'openai/gpt-6-sol-fast'
+  | 'openai/gpt-6.1-sol'
+  | 'openai/gpt-6.1-sol-fast'
   | 'openai/gpt-oss-120b'
   | 'openai/gpt-oss-20b'
   | 'openai/gpt-oss-safeguard-120b'

package/src/gateway-speech-model.ts

@@ -84,6 +84,7 @@ export class GatewaySpeechModel implements SpeechModelV4 {
       return {
         audio: responseBody.audio,
         warnings: (responseBody.warnings ?? []) as Array<SharedV4Warning>,
+        ...(responseBody.usage != null && { usage: responseBody.usage }),
         providerMetadata:
           responseBody.providerMetadata as SharedV4ProviderMetadata,
         response: {
@@ -140,6 +141,7 @@ const gatewaySpeechWarningSchema = z.discriminatedUnion('type', [
 const gatewaySpeechResponseSchema = z.object({
   audio: z.string(),
   warnings: z.array(gatewaySpeechWarningSchema).optional(),
+  usage: z.record(z.string(), z.json()).optional(),
   providerMetadata: z
     .record(z.string(), providerMetadataEntrySchema)
     .optional(),

package/src/gateway-transcription-model.ts

@@ -105,6 +105,7 @@ export class GatewayTranscriptionModel implements TranscriptionModelV4 {
         language: responseBody.language ?? undefined,
         durationInSeconds: responseBody.durationInSeconds ?? undefined,
         warnings: (responseBody.warnings ?? []) as Array<SharedV4Warning>,
+        ...(responseBody.usage != null && { usage: responseBody.usage }),
         providerMetadata:
           responseBody.providerMetadata as SharedV4ProviderMetadata,
         response: {
@@ -433,6 +434,7 @@ const gatewayTranscriptionResponseSchema = z.object({
   language: z.string().nullish(),
   durationInSeconds: z.number().nullish(),
   warnings: z.array(gatewayTranscriptionWarningSchema).optional(),
+  usage: z.record(z.string(), z.json()).optional(),
   providerMetadata: z
     .record(z.string(), providerMetadataEntrySchema)
     .optional(),

package/src/zod.ts

@@ -5,6 +5,7 @@ import {
   discriminatedUnion,
   enum as enumType,
   literal,
+  json,
   number,
   object,
   record,
@@ -22,6 +23,7 @@ export const z = {
   discriminatedUnion,
   enum: enumType,
   literal,
+  json,
   number,
   object,
   record,