@ai-sdk/gateway 4.0.0-beta.6 → 4.0.0-beta.61

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

@@ -5,7 +5,7 @@ export type GatewayEmbeddingModelId =
   | 'amazon/titan-embed-text-v2'
   | 'cohere/embed-v4.0'
   | 'google/gemini-embedding-001'
-  | 'google/gemini-embedding-2-preview'
+  | 'google/gemini-embedding-2'
   | 'google/text-embedding-005'
   | 'google/text-multilingual-embedding-002'
   | 'mistral/codestral-embed'

package/src/gateway-embedding-model.ts

@@ -1,6 +1,6 @@
 import type {
-  EmbeddingModelV3,
-  SharedV3ProviderMetadata,
+  EmbeddingModelV4,
+  SharedV4ProviderMetadata,
 } from '@ai-sdk/provider';
 import {
   combineHeaders,
@@ -9,6 +9,9 @@ import {
   lazySchema,
   postJsonToApi,
   resolve,
+  serializeModelOptions,
+  WORKFLOW_SERIALIZE,
+  WORKFLOW_DESERIALIZE,
   zodSchema,
   type Resolvable,
 } from '@ai-sdk/provider-utils';
@@ -17,17 +20,33 @@ import { asGatewayError } from './errors';
 import { parseAuthMethod } from './errors/parse-auth-method';
 import type { GatewayConfig } from './gateway-config';
 
-export class GatewayEmbeddingModel implements EmbeddingModelV3 {
-  readonly specificationVersion = 'v3';
+type GatewayEmbeddingConfig = GatewayConfig & {
+  provider: string;
+  o11yHeaders: Resolvable<Record<string, string>>;
+};
+
+export class GatewayEmbeddingModel implements EmbeddingModelV4 {
+  readonly specificationVersion = 'v4';
   readonly maxEmbeddingsPerCall = 2048;
   readonly supportsParallelCalls = true;
 
+  static [WORKFLOW_SERIALIZE](model: GatewayEmbeddingModel) {
+    return serializeModelOptions({
+      modelId: model.modelId,
+      config: model.config,
+    });
+  }
+
+  static [WORKFLOW_DESERIALIZE](options: {
+    modelId: string;
+    config: GatewayEmbeddingConfig;
+  }) {
+    return new GatewayEmbeddingModel(options.modelId, options.config);
+  }
+
   constructor(
     readonly modelId: string,
-    private readonly config: GatewayConfig & {
-      provider: string;
-      o11yHeaders: Resolvable<Record<string, string>>;
-    },
+    private readonly config: GatewayEmbeddingConfig,
   ) {}
 
   get provider(): string {
@@ -39,10 +58,12 @@ export class GatewayEmbeddingModel implements EmbeddingModelV3 {
     headers,
     abortSignal,
     providerOptions,
-  }: Parameters<EmbeddingModelV3['doEmbed']>[0]): Promise<
-    Awaited<ReturnType<EmbeddingModelV3['doEmbed']>>
+  }: Parameters<EmbeddingModelV4['doEmbed']>[0]): Promise<
+    Awaited<ReturnType<EmbeddingModelV4['doEmbed']>>
   > {
-    const resolvedHeaders = await resolve(this.config.headers());
+    const resolvedHeaders = this.config.headers
+      ? await resolve(this.config.headers)
+      : undefined;
     try {
       const {
         responseHeaders,
@@ -75,12 +96,15 @@ export class GatewayEmbeddingModel implements EmbeddingModelV3 {
         embeddings: responseBody.embeddings,
         usage: responseBody.usage ?? undefined,
         providerMetadata:
-          responseBody.providerMetadata as unknown as SharedV3ProviderMetadata,
+          responseBody.providerMetadata as unknown as SharedV4ProviderMetadata,
         response: { headers: responseHeaders, body: rawValue },
         warnings: [],
       };
     } catch (error) {
-      throw await asGatewayError(error, await parseAuthMethod(resolvedHeaders));
+      throw await asGatewayError(
+        error,
+        await parseAuthMethod(resolvedHeaders ?? {}),
+      );
     }
   }
 
@@ -90,7 +114,7 @@ export class GatewayEmbeddingModel implements EmbeddingModelV3 {
 
   private getModelConfigHeaders() {
     return {
-      'ai-embedding-model-specification-version': '3',
+      'ai-embedding-model-specification-version': '4',
       'ai-model-id': this.modelId,
     };
   }

package/src/gateway-fetch-metadata.ts

@@ -9,7 +9,11 @@ import {
 import { z } from 'zod/v4';
 import { asGatewayError } from './errors';
 import type { GatewayConfig } from './gateway-config';
-import type { GatewayLanguageModelEntry } from './gateway-model-entry';
+import {
+  KNOWN_MODEL_TYPES,
+  type GatewayLanguageModelEntry,
+  type KnownModelType,
+} from './gateway-model-entry';
 
 type GatewayFetchMetadataConfig = GatewayConfig;
 
@@ -31,7 +35,9 @@ export class GatewayFetchMetadata {
     try {
       const { value } = await getFromApi({
         url: `${this.config.baseURL}/config`,
-        headers: await resolve(this.config.headers()),
+        headers: this.config.headers
+          ? await resolve(this.config.headers)
+          : undefined,
         successfulResponseHandler: createJsonResponseHandler(
           gatewayAvailableModelsResponseSchema,
         ),
@@ -54,7 +60,9 @@ export class GatewayFetchMetadata {
 
       const { value } = await getFromApi({
         url: `${baseUrl.origin}/v1/credits`,
-        headers: await resolve(this.config.headers()),
+        headers: this.config.headers
+          ? await resolve(this.config.headers)
+          : undefined,
         successfulResponseHandler: createJsonResponseHandler(
           gatewayCreditsResponseSchema,
         ),
@@ -75,41 +83,47 @@ export class GatewayFetchMetadata {
 const gatewayAvailableModelsResponseSchema = lazySchema(() =>
   zodSchema(
     z.object({
-      models: z.array(
-        z.object({
-          id: z.string(),
-          name: z.string(),
-          description: z.string().nullish(),
-          pricing: z
-            .object({
-              input: z.string(),
-              output: z.string(),
-              input_cache_read: z.string().nullish(),
-              input_cache_write: z.string().nullish(),
-            })
-            .transform(
-              ({ input, output, input_cache_read, input_cache_write }) => ({
-                input,
-                output,
-                ...(input_cache_read
-                  ? { cachedInputTokens: input_cache_read }
-                  : {}),
-                ...(input_cache_write
-                  ? { cacheCreationInputTokens: input_cache_write }
-                  : {}),
-              }),
-            )
-            .nullish(),
-          specification: z.object({
-            specificationVersion: z.literal('v3'),
-            provider: z.string(),
-            modelId: z.string(),
+      models: z
+        .array(
+          z.object({
+            id: z.string(),
+            name: z.string(),
+            description: z.string().nullish(),
+            pricing: z
+              .object({
+                input: z.string(),
+                output: z.string(),
+                input_cache_read: z.string().nullish(),
+                input_cache_write: z.string().nullish(),
+              })
+              .transform(
+                ({ input, output, input_cache_read, input_cache_write }) => ({
+                  input,
+                  output,
+                  ...(input_cache_read
+                    ? { cachedInputTokens: input_cache_read }
+                    : {}),
+                  ...(input_cache_write
+                    ? { cacheCreationInputTokens: input_cache_write }
+                    : {}),
+                }),
+              )
+              .nullish(),
+            specification: z.object({
+              specificationVersion: z.literal('v4'),
+              provider: z.string(),
+              modelId: z.string(),
+            }),
+            modelType: z.string().nullish(),
           }),
-          modelType: z
-            .enum(['embedding', 'image', 'language', 'video'])
-            .nullish(),
-        }),
-      ),
+        )
+        .transform(models =>
+          models.filter(
+            (m): m is typeof m & { modelType?: KnownModelType | null } =>
+              m.modelType == null ||
+              KNOWN_MODEL_TYPES.includes(m.modelType as KnownModelType),
+          ),
+        ),
     }),
   ),
 );

package/src/gateway-generation-info.ts

@@ -0,0 +1,149 @@
+import {
+  createJsonErrorResponseHandler,
+  createJsonResponseHandler,
+  getFromApi,
+  lazySchema,
+  resolve,
+  zodSchema,
+} from '@ai-sdk/provider-utils';
+import { z } from 'zod/v4';
+import { asGatewayError } from './errors';
+import type { GatewayConfig } from './gateway-config';
+
+export interface GatewayGenerationInfoParams {
+  /** The generation ID to look up (format: gen_<ulid>) */
+  id: string;
+}
+
+export interface GatewayGenerationInfo {
+  /** The generation ID */
+  id: string;
+  /** Total cost in USD */
+  totalCost: number;
+  /** Upstream inference cost in USD (BYOK only) */
+  upstreamInferenceCost: number;
+  /** Usage cost in USD (same as totalCost) */
+  usage: number;
+  /** ISO 8601 timestamp when the generation was created */
+  createdAt: string;
+  /** Model identifier */
+  model: string;
+  /** Whether BYOK credentials were used */
+  isByok: boolean;
+  /** Provider that served this generation */
+  providerName: string;
+  /** Whether streaming was used */
+  streamed: boolean;
+  /** Finish reason (e.g. 'stop') */
+  finishReason: string;
+  /** Time to first token in milliseconds */
+  latency: number;
+  /** Total generation time in milliseconds */
+  generationTime: number;
+  /** Number of prompt tokens */
+  promptTokens: number;
+  /** Number of completion tokens */
+  completionTokens: number;
+  /** Reasoning tokens used */
+  reasoningTokens: number;
+  /** Cached tokens used */
+  cachedTokens: number;
+  /** Cache creation input tokens */
+  cacheCreationTokens: number;
+  /** Billable web search calls */
+  billableWebSearchCalls: number;
+}
+
+export class GatewayGenerationInfoFetcher {
+  constructor(private readonly config: GatewayConfig) {}
+
+  async getGenerationInfo(
+    params: GatewayGenerationInfoParams,
+  ): Promise<GatewayGenerationInfo> {
+    try {
+      const baseUrl = new URL(this.config.baseURL);
+
+      const { value } = await getFromApi({
+        url: `${baseUrl.origin}/v1/generation?id=${encodeURIComponent(params.id)}`,
+        headers: this.config.headers
+          ? await resolve(this.config.headers)
+          : undefined,
+        successfulResponseHandler: createJsonResponseHandler(
+          gatewayGenerationInfoResponseSchema,
+        ),
+        failedResponseHandler: createJsonErrorResponseHandler({
+          errorSchema: z.any(),
+          errorToMessage: data => data,
+        }),
+        fetch: this.config.fetch,
+      });
+
+      return value;
+    } catch (error) {
+      throw await asGatewayError(error);
+    }
+  }
+}
+
+const gatewayGenerationInfoResponseSchema = lazySchema(() =>
+  zodSchema(
+    z
+      .object({
+        data: z
+          .object({
+            id: z.string(),
+            total_cost: z.number(),
+            upstream_inference_cost: z.number(),
+            usage: z.number(),
+            created_at: z.string(),
+            model: z.string(),
+            is_byok: z.boolean(),
+            provider_name: z.string(),
+            streamed: z.boolean(),
+            finish_reason: z.string(),
+            latency: z.number(),
+            generation_time: z.number(),
+            native_tokens_prompt: z.number(),
+            native_tokens_completion: z.number(),
+            native_tokens_reasoning: z.number(),
+            native_tokens_cached: z.number(),
+            native_tokens_cache_creation: z.number(),
+            billable_web_search_calls: z.number(),
+          })
+          .transform(
+            ({
+              total_cost,
+              upstream_inference_cost,
+              created_at,
+              is_byok,
+              provider_name,
+              finish_reason,
+              generation_time,
+              native_tokens_prompt,
+              native_tokens_completion,
+              native_tokens_reasoning,
+              native_tokens_cached,
+              native_tokens_cache_creation,
+              billable_web_search_calls,
+              ...rest
+            }) => ({
+              ...rest,
+              totalCost: total_cost,
+              upstreamInferenceCost: upstream_inference_cost,
+              createdAt: created_at,
+              isByok: is_byok,
+              providerName: provider_name,
+              finishReason: finish_reason,
+              generationTime: generation_time,
+              promptTokens: native_tokens_prompt,
+              completionTokens: native_tokens_completion,
+              reasoningTokens: native_tokens_reasoning,
+              cachedTokens: native_tokens_cached,
+              cacheCreationTokens: native_tokens_cache_creation,
+              billableWebSearchCalls: billable_web_search_calls,
+            }),
+          ),
+      })
+      .transform(({ data }) => data),
+  ),
+);

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

@@ -1,15 +1,24 @@
 export type GatewayImageModelId =
+  | 'bfl/flux-2-flex'
+  | 'bfl/flux-2-klein-4b'
+  | 'bfl/flux-2-klein-9b'
+  | 'bfl/flux-2-max'
+  | 'bfl/flux-2-pro'
   | 'bfl/flux-kontext-max'
   | 'bfl/flux-kontext-pro'
   | 'bfl/flux-pro-1.0-fill'
   | 'bfl/flux-pro-1.1'
   | 'bfl/flux-pro-1.1-ultra'
+  | 'bytedance/seedream-4.0'
+  | 'bytedance/seedream-4.5'
+  | 'bytedance/seedream-5.0-lite'
   | 'google/imagen-4.0-fast-generate-001'
   | 'google/imagen-4.0-generate-001'
   | 'google/imagen-4.0-ultra-generate-001'
   | 'openai/gpt-image-1'
   | 'openai/gpt-image-1-mini'
   | 'openai/gpt-image-1.5'
+  | 'prodia/flux-fast-schnell'
   | 'recraft/recraft-v2'
   | 'recraft/recraft-v3'
   | 'recraft/recraft-v4'

package/src/gateway-image-model.ts

@@ -1,7 +1,7 @@
 import type {
-  ImageModelV3,
-  ImageModelV3File,
-  ImageModelV3ProviderMetadata,
+  ImageModelV4,
+  ImageModelV4File,
+  ImageModelV4ProviderMetadata,
 } from '@ai-sdk/provider';
 import {
   combineHeaders,
@@ -10,6 +10,9 @@ import {
   createJsonErrorResponseHandler,
   postJsonToApi,
   resolve,
+  serializeModelOptions,
+  WORKFLOW_SERIALIZE,
+  WORKFLOW_DESERIALIZE,
   type Resolvable,
 } from '@ai-sdk/provider-utils';
 import { z } from 'zod/v4';
@@ -17,17 +20,33 @@ import type { GatewayConfig } from './gateway-config';
 import { asGatewayError } from './errors';
 import { parseAuthMethod } from './errors/parse-auth-method';
 
-export class GatewayImageModel implements ImageModelV3 {
-  readonly specificationVersion = 'v3' as const;
+type GatewayImageConfig = GatewayConfig & {
+  provider: string;
+  o11yHeaders: Resolvable<Record<string, string>>;
+};
+
+export class GatewayImageModel implements ImageModelV4 {
+  readonly specificationVersion = 'v4' as const;
   // Set a very large number to prevent client-side splitting of requests
   readonly maxImagesPerCall = Number.MAX_SAFE_INTEGER;
 
+  static [WORKFLOW_SERIALIZE](model: GatewayImageModel) {
+    return serializeModelOptions({
+      modelId: model.modelId,
+      config: model.config,
+    });
+  }
+
+  static [WORKFLOW_DESERIALIZE](options: {
+    modelId: string;
+    config: GatewayImageConfig;
+  }) {
+    return new GatewayImageModel(options.modelId, options.config);
+  }
+
   constructor(
     readonly modelId: string,
-    private readonly config: GatewayConfig & {
-      provider: string;
-      o11yHeaders: Resolvable<Record<string, string>>;
-    },
+    private readonly config: GatewayImageConfig,
   ) {}
 
   get provider(): string {
@@ -45,16 +64,14 @@ export class GatewayImageModel implements ImageModelV3 {
     providerOptions,
     headers,
     abortSignal,
-  }: Parameters<ImageModelV3['doGenerate']>[0]): Promise<
-    Awaited<ReturnType<ImageModelV3['doGenerate']>>
+  }: Parameters<ImageModelV4['doGenerate']>[0]): Promise<
+    Awaited<ReturnType<ImageModelV4['doGenerate']>>
   > {
-    const resolvedHeaders = await resolve(this.config.headers());
+    const resolvedHeaders = this.config.headers
+      ? await resolve(this.config.headers)
+      : undefined;
     try {
-      const {
-        responseHeaders,
-        value: responseBody,
-        rawValue,
-      } = await postJsonToApi({
+      const { responseHeaders, value: responseBody } = await postJsonToApi({
         url: this.getUrl(),
         headers: combineHeaders(
           resolvedHeaders,
@@ -89,7 +106,7 @@ export class GatewayImageModel implements ImageModelV3 {
         images: responseBody.images, // Always base64 strings from server
         warnings: responseBody.warnings ?? [],
         providerMetadata:
-          responseBody.providerMetadata as ImageModelV3ProviderMetadata,
+          responseBody.providerMetadata as ImageModelV4ProviderMetadata,
         response: {
           timestamp: new Date(),
           modelId: this.modelId,
@@ -104,7 +121,10 @@ export class GatewayImageModel implements ImageModelV3 {
         }),
       };
     } catch (error) {
-      throw await asGatewayError(error, await parseAuthMethod(resolvedHeaders));
+      throw await asGatewayError(
+        error,
+        await parseAuthMethod(resolvedHeaders ?? {}),
+      );
     }
   }
 
@@ -114,13 +134,13 @@ export class GatewayImageModel implements ImageModelV3 {
 
   private getModelConfigHeaders() {
     return {
-      'ai-image-model-specification-version': '3',
+      'ai-image-model-specification-version': '4',
       'ai-model-id': this.modelId,
     };
   }
 }
 
-function maybeEncodeImageFile(file: ImageModelV3File) {
+function maybeEncodeImageFile(file: ImageModelV4File) {
   if (file.type === 'file' && file.data instanceof Uint8Array) {
     return {
       ...file,

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

@@ -17,25 +17,25 @@ export type GatewayModelId =
   | 'alibaba/qwen3-vl-thinking'
   | 'alibaba/qwen3.5-flash'
   | 'alibaba/qwen3.5-plus'
+  | 'alibaba/qwen3.6-plus'
   | 'amazon/nova-2-lite'
   | 'amazon/nova-lite'
   | 'amazon/nova-micro'
   | 'amazon/nova-pro'
   | 'anthropic/claude-3-haiku'
-  | 'anthropic/claude-3-opus'
   | 'anthropic/claude-3.5-haiku'
-  | 'anthropic/claude-3.5-sonnet'
-  | 'anthropic/claude-3.5-sonnet-20240620'
   | 'anthropic/claude-3.7-sonnet'
   | 'anthropic/claude-haiku-4.5'
   | 'anthropic/claude-opus-4'
   | 'anthropic/claude-opus-4.1'
   | 'anthropic/claude-opus-4.5'
   | 'anthropic/claude-opus-4.6'
+  | 'anthropic/claude-opus-4.7'
   | 'anthropic/claude-sonnet-4'
   | 'anthropic/claude-sonnet-4.5'
   | 'anthropic/claude-sonnet-4.6'
   | 'arcee-ai/trinity-large-preview'
+  | 'arcee-ai/trinity-large-thinking'
   | 'arcee-ai/trinity-mini'
   | 'bytedance/seed-1.6'
   | 'bytedance/seed-1.8'
@@ -51,8 +51,6 @@ export type GatewayModelId =
   | 'google/gemini-2.5-flash'
   | 'google/gemini-2.5-flash-image'
   | 'google/gemini-2.5-flash-lite'
-  | 'google/gemini-2.5-flash-lite-preview-09-2025'
-  | 'google/gemini-2.5-flash-preview-09-2025'
   | 'google/gemini-2.5-pro'
   | 'google/gemini-3-flash'
   | 'google/gemini-3-pro-image'
@@ -60,11 +58,14 @@ export type GatewayModelId =
   | 'google/gemini-3.1-flash-image-preview'
   | 'google/gemini-3.1-flash-lite-preview'
   | 'google/gemini-3.1-pro-preview'
+  | 'google/gemma-4-26b-a4b-it'
+  | 'google/gemma-4-31b-it'
   | 'inception/mercury-2'
   | 'inception/mercury-coder-small'
   | 'kwaipilot/kat-coder-pro-v1'
+  | 'kwaipilot/kat-coder-pro-v2'
   | 'meituan/longcat-flash-chat'
-  | 'meituan/longcat-flash-thinking'
+  | 'meituan/longcat-flash-thinking-2601'
   | 'meta/llama-3.1-70b'
   | 'meta/llama-3.1-8b'
   | 'meta/llama-3.2-11b'
@@ -79,6 +80,8 @@ export type GatewayModelId =
   | 'minimax/minimax-m2.1-lightning'
   | 'minimax/minimax-m2.5'
   | 'minimax/minimax-m2.5-highspeed'
+  | 'minimax/minimax-m2.7'
+  | 'minimax/minimax-m2.7-highspeed'
   | 'mistral/codestral'
   | 'mistral/devstral-2'
   | 'mistral/devstral-small'
@@ -104,9 +107,9 @@ export type GatewayModelId =
   | 'morph/morph-v3-fast'
   | 'morph/morph-v3-large'
   | 'nvidia/nemotron-3-nano-30b-a3b'
+  | 'nvidia/nemotron-3-super-120b-a12b'
   | 'nvidia/nemotron-nano-12b-v2-vl'
   | 'nvidia/nemotron-nano-9b-v2'
-  | 'openai/codex-mini'
   | 'openai/gpt-3.5-turbo'
   | 'openai/gpt-3.5-turbo-instruct'
   | 'openai/gpt-4-turbo'
@@ -134,6 +137,8 @@ export type GatewayModelId =
   | 'openai/gpt-5.3-chat'
   | 'openai/gpt-5.3-codex'
   | 'openai/gpt-5.4'
+  | 'openai/gpt-5.4-mini'
+  | 'openai/gpt-5.4-nano'
   | 'openai/gpt-5.4-pro'
   | 'openai/gpt-oss-120b'
   | 'openai/gpt-oss-20b'
@@ -146,11 +151,8 @@ export type GatewayModelId =
   | 'openai/o4-mini'
   | 'perplexity/sonar'
   | 'perplexity/sonar-pro'
-  | 'perplexity/sonar-reasoning'
   | 'perplexity/sonar-reasoning-pro'
   | 'prime-intellect/intellect-3'
-  | 'vercel/v0-1.0-md'
-  | 'vercel/v0-1.5-md'
   | 'xai/grok-3'
   | 'xai/grok-3-fast'
   | 'xai/grok-3-mini'
@@ -160,8 +162,15 @@ export type GatewayModelId =
   | 'xai/grok-4-fast-reasoning'
   | 'xai/grok-4.1-fast-non-reasoning'
   | 'xai/grok-4.1-fast-reasoning'
+  | 'xai/grok-4.20-multi-agent'
+  | 'xai/grok-4.20-multi-agent-beta'
+  | 'xai/grok-4.20-non-reasoning'
+  | 'xai/grok-4.20-non-reasoning-beta'
+  | 'xai/grok-4.20-reasoning'
+  | 'xai/grok-4.20-reasoning-beta'
   | 'xai/grok-code-fast-1'
   | 'xiaomi/mimo-v2-flash'
+  | 'xiaomi/mimo-v2-pro'
   | 'zai/glm-4.5'
   | 'zai/glm-4.5-air'
   | 'zai/glm-4.5v'
@@ -172,4 +181,7 @@ export type GatewayModelId =
   | 'zai/glm-4.7-flash'
   | 'zai/glm-4.7-flashx'
   | 'zai/glm-5'
+  | 'zai/glm-5-turbo'
+  | 'zai/glm-5.1'
+  | 'zai/glm-5v-turbo'
   | (string & {});