@ai-sdk/google 4.0.79 → 4.0.80

package/docs/15-google.mdx

@@ -923,14 +923,14 @@ The `vertexRagStore` tool accepts the following configuration options:
 
 ### Image Outputs
 
-Gemini models with image generation capabilities (e.g. `gemini-2.5-flash-image`) support generating images as part of a multimodal response. Images are exposed as files in the response.
+Gemini models with image generation capabilities (e.g. `gemini-3.1-flash-image-preview`) support generating images as part of a multimodal response. Images are exposed as files in the response.
 
 ```ts
 import { google } from '@ai-sdk/google';
 import { generateText } from 'ai';
 
 const result = await generateText({
-  model: google('gemini-2.5-flash-image'),
+  model: google('gemini-3.1-flash-image-preview'),
   prompt:
     'Create a picture of a nano banana dish in a fancy restaurant with a Gemini theme',
 });
@@ -1088,7 +1088,7 @@ You can pass a `webhookUrl` to receive a notification when the batch reaches a t
 import { google } from '@ai-sdk/google';
 import { experimental_startBatch as startBatch } from 'ai';
 
-const model = 'gemini-3.6-flash';
+const model = 'gemini-3.8-flash';
 
 const batch = await startBatch({
   provider: google,
@@ -1578,7 +1578,7 @@ import { google, type GoogleInteractionsVideoOptions } from '@ai-sdk/google';
 import { generateText } from 'ai';
 
 const result = await generateText({
-  model: google.interactions('gemini-3.7-flash'),
+  model: google.interactions('gemini-3.8-flash'),
   messages: [
     {
       role: 'user',
@@ -1778,7 +1778,7 @@ import {
 import { generateText } from 'ai';
 
 const result = await generateText({
-  model: google.interactions('gemini-2.5-flash-image'),
+  model: google.interactions('gemini-3.1-flash-image-preview'),
   prompt:
     'Tell me a three sentence bedtime story about a unicorn, accompanied by a suitable illustration.',
   providerOptions: {
@@ -2042,7 +2042,7 @@ You can create models that call the [Google Generative AI embeddings API](https:
 using the `.embedding()` factory method.
 
 ```ts
-const model = google.embedding('gemini-embedding-001');
+const model = google.embedding('gemini-embedding-2');
 ```
 
 The Google provider sends API calls to the right endpoint based on the type of embedding:
@@ -2078,7 +2078,7 @@ import { google, type GoogleEmbeddingModelOptions } from '@ai-sdk/google';
 import { embedMany } from 'ai';
 
 const { embeddings } = await embedMany({
-  model: google.embedding('gemini-embedding-2-preview'),
+  model: google.embedding('gemini-embedding-2'),
   values: ['sunny day at the beach', 'rainy afternoon in the city'],
   providerOptions: {
     google: {
@@ -2132,14 +2132,14 @@ output capabilities through the `:generateContent` API.
 
 ### Gemini Image Models
 
-[Gemini image models](https://ai.google.dev/gemini-api/docs/image-generation) (e.g. `gemini-2.5-flash-image`) are technically multimodal output language models, but they can be used with the `generateImage()` function for a simpler image generation experience. Internally, the provider calls the language model API with `responseModalities: ['IMAGE']`.
+[Gemini image models](https://ai.google.dev/gemini-api/docs/image-generation) (e.g. `gemini-3.1-flash-image-preview`) are technically multimodal output language models, but they can be used with the `generateImage()` function for a simpler image generation experience. Internally, the provider calls the language model API with `responseModalities: ['IMAGE']`.
 
 ```ts
 import { google } from '@ai-sdk/google';
 import { generateImage } from 'ai';
 
 const { image } = await generateImage({
-  model: google.image('gemini-2.5-flash-image'),
+  model: google.image('gemini-3.1-flash-image-preview'),
   prompt: 'A photorealistic image of a cat wearing a wizard hat',
   aspectRatio: '1:1',
 });
@@ -2155,7 +2155,7 @@ import fs from 'node:fs';
 const sourceImage = fs.readFileSync('./cat.png');
 
 const { image } = await generateImage({
-  model: google.image('gemini-2.5-flash-image'),
+  model: google.image('gemini-3.1-flash-image-preview'),
   prompt: {
     text: 'Add a small wizard hat to this cat',
     images: [sourceImage],
@@ -2170,7 +2170,7 @@ import { google } from '@ai-sdk/google';
 import { generateImage } from 'ai';
 
 const { image } = await generateImage({
-  model: google.image('gemini-2.5-flash-image'),
+  model: google.image('gemini-3.1-flash-image-preview'),
   prompt: {
     text: 'Add a small wizard hat to this cat',
     images: ['https://example.com/cat.png'],

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ai-sdk/google",
-  "version": "4.0.79",
+  "version": "4.0.80",
   "type": "module",
   "license": "Apache-2.0",
   "sideEffects": false,

package/src/convert-to-google-messages.ts

@@ -1,5 +1,6 @@
 import {
   UnsupportedFunctionalityError,
+  type JSONValue,
   type LanguageModelV4Prompt,
   type LanguageModelV4ToolResultOutput,
   type SharedV4Warning,
@@ -67,6 +68,30 @@ function convertUrlToolResultPart(
   };
 }
 
+function containsJSONSchemaReference(value: JSONValue | undefined): boolean {
+  if (Array.isArray(value)) {
+    return value.some(containsJSONSchemaReference);
+  }
+
+  if (typeof value !== 'object' || value === null) {
+    return false;
+  }
+
+  return Object.entries(value).some(
+    ([key, nestedValue]) =>
+      key === '$ref' || containsJSONSchemaReference(nestedValue),
+  );
+}
+
+function serializeFunctionResponseContent(
+  value: JSONValue,
+): JSONValue | string {
+  // Google reserves { $ref: displayName } in structured function responses for
+  // multimodal parts. This conflicts with JSON Schema $ref, so serialize the
+  // result to preserve it without triggering Google's reference handling.
+  return containsJSONSchemaReference(value) ? JSON.stringify(value) : value;
+}
+
 /*
  * Appends tool result content parts to the message using the functionResponse
  * format with support for multimodal parts (e.g. inline images/files alongside
@@ -654,7 +679,7 @@ export function convertToGoogleMessages(
                   content:
                     output.type === 'execution-denied'
                       ? (output.reason ?? 'Tool call execution denied.')
-                      : output.value,
+                      : serializeFunctionResponseContent(output.value),
                 },
               },
             });