@ai-sdk/openai 1.3.8 → 1.3.10

package/CHANGELOG.md

@@ -1,5 +1,22 @@
 # @ai-sdk/openai
 
+## 1.3.10
+
+### Patch Changes
+
+- dbe53e7: adding support for gpt-4o-search-preview and handling unsupported parameters
+- 84ffaba: fix: propagate openai transcription fixes
+
+## 1.3.9
+
+### Patch Changes
+
+- 013faa8: core (ai): change transcription model mimeType to mediaType
+- 013faa8: fix (provider/openai): increase transcription model resilience
+- Updated dependencies [013faa8]
+  - @ai-sdk/provider@1.1.2
+  - @ai-sdk/provider-utils@2.2.6
+
 ## 1.3.8
 
 ### Patch Changes

package/dist/index.d.mts

@@ -2,7 +2,7 @@ import { LanguageModelV1, ProviderV1, EmbeddingModelV1, ImageModelV1, Transcript
 import { FetchFunction } from '@ai-sdk/provider-utils';
 import { z } from 'zod';
 
-type OpenAIChatModelId = 'o1' | 'o1-2024-12-17' | 'o1-mini' | 'o1-mini-2024-09-12' | 'o1-preview' | 'o1-preview-2024-09-12' | 'o3-mini' | 'o3-mini-2025-01-31' | 'gpt-4o' | 'gpt-4o-2024-05-13' | 'gpt-4o-2024-08-06' | 'gpt-4o-2024-11-20' | 'gpt-4o-audio-preview' | 'gpt-4o-audio-preview-2024-10-01' | 'gpt-4o-audio-preview-2024-12-17' | 'gpt-4o-mini' | 'gpt-4o-mini-2024-07-18' | 'gpt-4-turbo' | 'gpt-4-turbo-2024-04-09' | 'gpt-4-turbo-preview' | 'gpt-4-0125-preview' | 'gpt-4-1106-preview' | 'gpt-4' | 'gpt-4-0613' | 'gpt-4.5-preview' | 'gpt-4.5-preview-2025-02-27' | 'gpt-3.5-turbo-0125' | 'gpt-3.5-turbo' | 'gpt-3.5-turbo-1106' | 'chatgpt-4o-latest' | (string & {});
+type OpenAIChatModelId = 'o1' | 'o1-2024-12-17' | 'o1-mini' | 'o1-mini-2024-09-12' | 'o1-preview' | 'o1-preview-2024-09-12' | 'o3-mini' | 'o3-mini-2025-01-31' | 'gpt-4o' | 'gpt-4o-2024-05-13' | 'gpt-4o-2024-08-06' | 'gpt-4o-2024-11-20' | 'gpt-4o-audio-preview' | 'gpt-4o-audio-preview-2024-10-01' | 'gpt-4o-audio-preview-2024-12-17' | 'gpt-4o-search-preview' | 'gpt-4o-search-preview-2025-03-11' | 'gpt-4o-mini' | 'gpt-4o-mini-2024-07-18' | 'gpt-4-turbo' | 'gpt-4-turbo-2024-04-09' | 'gpt-4-turbo-preview' | 'gpt-4-0125-preview' | 'gpt-4-1106-preview' | 'gpt-4' | 'gpt-4-0613' | 'gpt-4.5-preview' | 'gpt-4.5-preview-2025-02-27' | 'gpt-3.5-turbo-0125' | 'gpt-3.5-turbo' | 'gpt-3.5-turbo-1106' | 'chatgpt-4o-latest' | (string & {});
 interface OpenAIChatSettings {
     /**
   Modify the likelihood of specified tokens appearing in the completion.

package/dist/index.d.ts

@@ -2,7 +2,7 @@ import { LanguageModelV1, ProviderV1, EmbeddingModelV1, ImageModelV1, Transcript
 import { FetchFunction } from '@ai-sdk/provider-utils';
 import { z } from 'zod';
 
-type OpenAIChatModelId = 'o1' | 'o1-2024-12-17' | 'o1-mini' | 'o1-mini-2024-09-12' | 'o1-preview' | 'o1-preview-2024-09-12' | 'o3-mini' | 'o3-mini-2025-01-31' | 'gpt-4o' | 'gpt-4o-2024-05-13' | 'gpt-4o-2024-08-06' | 'gpt-4o-2024-11-20' | 'gpt-4o-audio-preview' | 'gpt-4o-audio-preview-2024-10-01' | 'gpt-4o-audio-preview-2024-12-17' | 'gpt-4o-mini' | 'gpt-4o-mini-2024-07-18' | 'gpt-4-turbo' | 'gpt-4-turbo-2024-04-09' | 'gpt-4-turbo-preview' | 'gpt-4-0125-preview' | 'gpt-4-1106-preview' | 'gpt-4' | 'gpt-4-0613' | 'gpt-4.5-preview' | 'gpt-4.5-preview-2025-02-27' | 'gpt-3.5-turbo-0125' | 'gpt-3.5-turbo' | 'gpt-3.5-turbo-1106' | 'chatgpt-4o-latest' | (string & {});
+type OpenAIChatModelId = 'o1' | 'o1-2024-12-17' | 'o1-mini' | 'o1-mini-2024-09-12' | 'o1-preview' | 'o1-preview-2024-09-12' | 'o3-mini' | 'o3-mini-2025-01-31' | 'gpt-4o' | 'gpt-4o-2024-05-13' | 'gpt-4o-2024-08-06' | 'gpt-4o-2024-11-20' | 'gpt-4o-audio-preview' | 'gpt-4o-audio-preview-2024-10-01' | 'gpt-4o-audio-preview-2024-12-17' | 'gpt-4o-search-preview' | 'gpt-4o-search-preview-2025-03-11' | 'gpt-4o-mini' | 'gpt-4o-mini-2024-07-18' | 'gpt-4-turbo' | 'gpt-4-turbo-2024-04-09' | 'gpt-4-turbo-preview' | 'gpt-4-0125-preview' | 'gpt-4-1106-preview' | 'gpt-4' | 'gpt-4-0613' | 'gpt-4.5-preview' | 'gpt-4.5-preview-2025-02-27' | 'gpt-3.5-turbo-0125' | 'gpt-3.5-turbo' | 'gpt-3.5-turbo-1106' | 'chatgpt-4o-latest' | (string & {});
 interface OpenAIChatSettings {
     /**
   Modify the likelihood of specified tokens appearing in the completion.

package/dist/index.js

@@ -536,6 +536,15 @@ var OpenAIChatLanguageModel = class {
         }
         baseArgs.max_tokens = void 0;
       }
+    } else if (this.modelId.startsWith("gpt-4o-search-preview")) {
+      if (baseArgs.temperature != null) {
+        baseArgs.temperature = void 0;
+        warnings.push({
+          type: "unsupported-setting",
+          setting: "temperature",
+          details: "temperature is not supported for the gpt-4o-search-preview model and has been removed."
+        });
+      }
     }
     switch (type) {
       case "regular": {
@@ -1605,17 +1614,11 @@ var openaiImageResponseSchema = import_zod5.z.object({
 var import_provider_utils7 = require("@ai-sdk/provider-utils");
 var import_zod6 = require("zod");
 var OpenAIProviderOptionsSchema = import_zod6.z.object({
-  include: import_zod6.z.array(import_zod6.z.string()).optional().describe(
-    "Additional information to include in the transcription response."
-  ),
-  language: import_zod6.z.string().optional().describe("The language of the input audio in ISO-639-1 format."),
-  prompt: import_zod6.z.string().optional().describe(
-    "An optional text to guide the model's style or continue a previous audio segment."
-  ),
-  temperature: import_zod6.z.number().min(0).max(1).optional().default(0).describe("The sampling temperature, between 0 and 1."),
-  timestampGranularities: import_zod6.z.array(import_zod6.z.enum(["word", "segment"])).optional().default(["segment"]).describe(
-    "The timestamp granularities to populate for this transcription."
-  )
+  include: import_zod6.z.array(import_zod6.z.string()).nullish(),
+  language: import_zod6.z.string().nullish(),
+  prompt: import_zod6.z.string().nullish(),
+  temperature: import_zod6.z.number().min(0).max(1).nullish().default(0),
+  timestampGranularities: import_zod6.z.array(import_zod6.z.enum(["word", "segment"])).nullish().default(["segment"])
 });
 var languageMap = {
   afrikaans: "af",
@@ -1687,9 +1690,10 @@ var OpenAITranscriptionModel = class {
   }
   getArgs({
     audio,
-    mimeType,
+    mediaType,
     providerOptions
   }) {
+    var _a, _b, _c, _d, _e;
     const warnings = [];
     const openAIOptions = (0, import_provider_utils7.parseProviderOptions)({
       provider: "openai",
@@ -1699,19 +1703,19 @@ var OpenAITranscriptionModel = class {
     const formData = new FormData();
     const blob = audio instanceof Uint8Array ? new Blob([audio]) : new Blob([(0, import_provider_utils7.convertBase64ToUint8Array)(audio)]);
     formData.append("model", this.modelId);
-    formData.append("file", new File([blob], "audio", { type: mimeType }));
+    formData.append("file", new File([blob], "audio", { type: mediaType }));
     if (openAIOptions) {
       const transcriptionModelOptions = {
-        include: openAIOptions.include,
-        language: openAIOptions.language,
-        prompt: openAIOptions.prompt,
-        temperature: openAIOptions.temperature,
-        timestamp_granularities: openAIOptions.timestampGranularities
+        include: (_a = openAIOptions.include) != null ? _a : void 0,
+        language: (_b = openAIOptions.language) != null ? _b : void 0,
+        prompt: (_c = openAIOptions.prompt) != null ? _c : void 0,
+        temperature: (_d = openAIOptions.temperature) != null ? _d : void 0,
+        timestamp_granularities: (_e = openAIOptions.timestampGranularities) != null ? _e : void 0
       };
       for (const key in transcriptionModelOptions) {
         const value = transcriptionModelOptions[key];
         if (value !== void 0) {
-          formData.append(key, value);
+          formData.append(key, String(value));
         }
       }
     }
@@ -1721,10 +1725,14 @@ var OpenAITranscriptionModel = class {
     };
   }
   async doGenerate(options) {
-    var _a, _b, _c;
+    var _a, _b, _c, _d, _e, _f;
     const currentDate = (_c = (_b = (_a = this.config._internal) == null ? void 0 : _a.currentDate) == null ? void 0 : _b.call(_a)) != null ? _c : /* @__PURE__ */ new Date();
     const { formData, warnings } = this.getArgs(options);
-    const { value: response, responseHeaders } = await (0, import_provider_utils7.postFormDataToApi)({
+    const {
+      value: response,
+      responseHeaders,
+      rawValue: rawResponse
+    } = await (0, import_provider_utils7.postFormDataToApi)({
       url: this.config.url({
         path: "/audio/transcriptions",
         modelId: this.modelId
@@ -1738,46 +1746,37 @@ var OpenAITranscriptionModel = class {
       abortSignal: options.abortSignal,
       fetch: this.config.fetch
     });
-    let language;
-    if (response.language && response.language in languageMap) {
-      language = languageMap[response.language];
-    }
+    const language = response.language != null && response.language in languageMap ? languageMap[response.language] : void 0;
     return {
       text: response.text,
-      segments: response.words.map((word) => ({
+      segments: (_e = (_d = response.words) == null ? void 0 : _d.map((word) => ({
         text: word.word,
         startSecond: word.start,
         endSecond: word.end
-      })),
+      }))) != null ? _e : [],
       language,
-      durationInSeconds: response.duration,
+      durationInSeconds: (_f = response.duration) != null ? _f : void 0,
       warnings,
       response: {
         timestamp: currentDate,
         modelId: this.modelId,
         headers: responseHeaders,
-        body: response
-      },
-      // When using format `verbose_json` on `whisper-1`, OpenAI includes the things like `task` and enhanced `segments` information.
-      providerMetadata: {
-        openai: {
-          transcript: response
-        }
+        body: rawResponse
       }
     };
   }
 };
 var openaiTranscriptionResponseSchema = import_zod6.z.object({
   text: import_zod6.z.string(),
-  language: import_zod6.z.string().optional(),
-  duration: import_zod6.z.number().optional(),
+  language: import_zod6.z.string().nullish(),
+  duration: import_zod6.z.number().nullish(),
   words: import_zod6.z.array(
     import_zod6.z.object({
       word: import_zod6.z.string(),
       start: import_zod6.z.number(),
       end: import_zod6.z.number()
     })
-  )
+  ).nullish()
 });
 
 // src/responses/openai-responses-language-model.ts