diff --git a/.changeset/khaki-ladybugs-kiss.md b/.changeset/khaki-ladybugs-kiss.md
new file mode 100644
index 00000000000..a860a7d299f
--- /dev/null
+++ b/.changeset/khaki-ladybugs-kiss.md
@@ -0,0 +1,5 @@
+---
+'@firebase/vertexai': minor
+---
+
+Add an `AbortSignal` to request options for generation methods, allowing them to be aborted.
diff --git a/common/api-review/vertexai.api.md b/common/api-review/vertexai.api.md
index e7f00c2f4e0..d23e7e6ccb0 100644
--- a/common/api-review/vertexai.api.md
+++ b/common/api-review/vertexai.api.md
@@ -50,8 +50,8 @@ export class ChatSession {
     params?: StartChatParams | undefined;
     // (undocumented)
     requestOptions?: RequestOptions | undefined;
-    sendMessage(request: string | Array<string | Part>): Promise<GenerateContentResult>;
-    sendMessageStream(request: string | Array<string | Part>): Promise<GenerateContentStreamResult>;
+    sendMessage(request: string | Array<string | Part>, singleRequestOptions?: SingleRequestOptions): Promise<GenerateContentResult>;
+    sendMessageStream(request: string | Array<string | Part>, singleRequestOptions?: SingleRequestOptions): Promise<GenerateContentStreamResult>;
     }
 
 // @public
@@ -325,9 +325,9 @@ export interface GenerativeContentBlob {
 // @public
 export class GenerativeModel extends VertexAIModel {
     constructor(vertexAI: VertexAI, modelParams: ModelParams, requestOptions?: RequestOptions);
-    countTokens(request: CountTokensRequest | string | Array<string | Part>): Promise<CountTokensResponse>;
-    generateContent(request: GenerateContentRequest | string | Array<string | Part>): Promise<GenerateContentResult>;
-    generateContentStream(request: GenerateContentRequest | string | Array<string | Part>): Promise<GenerateContentStreamResult>;
+    countTokens(request: CountTokensRequest | string | Array<string | Part>, singleRequestOptions?: SingleRequestOptions): Promise<CountTokensResponse>;
+    generateContent(request: GenerateContentRequest | string | Array<string | Part>, singleRequestOptions?: SingleRequestOptions): Promise<GenerateContentResult>;
+    generateContentStream(request: GenerateContentRequest | string | Array<string | Part>, singleRequestOptions?: SingleRequestOptions): Promise<GenerateContentStreamResult>;
     // (undocumented)
     generationConfig: GenerationConfig;
     // (undocumented)
@@ -463,9 +463,9 @@ export interface ImagenInlineImage {
 // @beta
 export class ImagenModel extends VertexAIModel {
     constructor(vertexAI: VertexAI, modelParams: ImagenModelParams, requestOptions?: RequestOptions | undefined);
-    generateImages(prompt: string): Promise<ImagenGenerationResponse<ImagenInlineImage>>;
+    generateImages(prompt: string, singleRequestOptions?: SingleRequestOptions): Promise<ImagenGenerationResponse<ImagenInlineImage>>;
     // @internal
-    generateImagesGCS(prompt: string, gcsURI: string): Promise<ImagenGenerationResponse<ImagenGCSImage>>;
+    generateImagesGCS(prompt: string, gcsURI: string, singleRequestOptions?: SingleRequestOptions): Promise<ImagenGenerationResponse<ImagenGCSImage>>;
     generationConfig?: ImagenGenerationConfig;
     // (undocumented)
     requestOptions?: RequestOptions | undefined;
@@ -719,6 +719,11 @@ export interface Segment {
     startIndex: number;
 }
 
+// @public
+export interface SingleRequestOptions extends RequestOptions {
+    signal?: AbortSignal;
+}
+
 // @public
 export interface StartChatParams extends BaseParams {
     // (undocumented)
diff --git a/docs-devsite/_toc.yaml b/docs-devsite/_toc.yaml
index 665222edb9d..0f5d51acec6 100644
--- a/docs-devsite/_toc.yaml
+++ b/docs-devsite/_toc.yaml
@@ -588,6 +588,8 @@ toc:
     path: /docs/reference/js/vertexai.schemashared.md
   - title: Segment
     path: /docs/reference/js/vertexai.segment.md
+  - title: SingleRequestOptions
+    path: /docs/reference/js/vertexai.singlerequestoptions.md
   - title: StartChatParams
     path: /docs/reference/js/vertexai.startchatparams.md
   - title: StringSchema
diff --git a/docs-devsite/vertexai.chatsession.md b/docs-devsite/vertexai.chatsession.md
index cc5a75ace16..5fa4a6f48a0 100644
--- a/docs-devsite/vertexai.chatsession.md
+++ b/docs-devsite/vertexai.chatsession.md
@@ -37,8 +37,8 @@ export declare class ChatSession
 |  Method | Modifiers | Description |
 |  --- | --- | --- |
 |  [getHistory()](./vertexai.chatsession.md#chatsessiongethistory) |  | Gets the chat history so far. Blocked prompts are not added to history. Neither blocked candidates nor the prompts that generated them are added to history. |
-|  [sendMessage(request)](./vertexai.chatsession.md#chatsessionsendmessage) |  | Sends a chat message and receives a non-streaming <code>[GenerateContentResult](./vertexai.generatecontentresult.md#generatecontentresult_interface)</code> |
-|  [sendMessageStream(request)](./vertexai.chatsession.md#chatsessionsendmessagestream) |  | Sends a chat message and receives the response as a <code>[GenerateContentStreamResult](./vertexai.generatecontentstreamresult.md#generatecontentstreamresult_interface)</code> containing an iterable stream and a response promise. |
+|  [sendMessage(request, singleRequestOptions)](./vertexai.chatsession.md#chatsessionsendmessage) |  | Sends a chat message and receives a non-streaming <code>[GenerateContentResult](./vertexai.generatecontentresult.md#generatecontentresult_interface)</code> |
+|  [sendMessageStream(request, singleRequestOptions)](./vertexai.chatsession.md#chatsessionsendmessagestream) |  | Sends a chat message and receives the response as a <code>[GenerateContentStreamResult](./vertexai.generatecontentstreamresult.md#generatecontentstreamresult_interface)</code> containing an iterable stream and a response promise. |
 
 ## ChatSession.(constructor)
 
@@ -103,7 +103,7 @@ Sends a chat message and receives a non-streaming <code>[GenerateContentResult](
 <b>Signature:</b>
 
 ```typescript
-sendMessage(request: string | Array<string | Part>): Promise<GenerateContentResult>;
+sendMessage(request: string | Array<string | Part>, singleRequestOptions?: SingleRequestOptions): Promise<GenerateContentResult>;
 ```
 
 #### Parameters
@@ -111,6 +111,7 @@ sendMessage(request: string | Array<string | Part>): Promise<GenerateContentResu
 |  Parameter | Type | Description |
 |  --- | --- | --- |
 |  request | string \| Array&lt;string \| [Part](./vertexai.md#part)<!-- -->&gt; |  |
+|  singleRequestOptions | [SingleRequestOptions](./vertexai.singlerequestoptions.md#singlerequestoptions_interface) |  |
 
 <b>Returns:</b>
 
@@ -123,7 +124,7 @@ Sends a chat message and receives the response as a <code>[GenerateContentStream
 <b>Signature:</b>
 
 ```typescript
-sendMessageStream(request: string | Array<string | Part>): Promise<GenerateContentStreamResult>;
+sendMessageStream(request: string | Array<string | Part>, singleRequestOptions?: SingleRequestOptions): Promise<GenerateContentStreamResult>;
 ```
 
 #### Parameters
@@ -131,6 +132,7 @@ sendMessageStream(request: string | Array<string | Part>): Promise<GenerateConte
 |  Parameter | Type | Description |
 |  --- | --- | --- |
 |  request | string \| Array&lt;string \| [Part](./vertexai.md#part)<!-- -->&gt; |  |
+|  singleRequestOptions | [SingleRequestOptions](./vertexai.singlerequestoptions.md#singlerequestoptions_interface) |  |
 
 <b>Returns:</b>
 
diff --git a/docs-devsite/vertexai.generativemodel.md b/docs-devsite/vertexai.generativemodel.md
index b734e241e78..6d691ed9e5a 100644
--- a/docs-devsite/vertexai.generativemodel.md
+++ b/docs-devsite/vertexai.generativemodel.md
@@ -40,9 +40,9 @@ export declare class GenerativeModel extends VertexAIModel
 
 |  Method | Modifiers | Description |
 |  --- | --- | --- |
-|  [countTokens(request)](./vertexai.generativemodel.md#generativemodelcounttokens) |  | Counts the tokens in the provided request. |
-|  [generateContent(request)](./vertexai.generativemodel.md#generativemodelgeneratecontent) |  | Makes a single non-streaming call to the model and returns an object containing a single <code>[GenerateContentResponse](./vertexai.generatecontentresponse.md#generatecontentresponse_interface)</code>. |
-|  [generateContentStream(request)](./vertexai.generativemodel.md#generativemodelgeneratecontentstream) |  | Makes a single streaming call to the model and returns an object containing an iterable stream that iterates over all chunks in the streaming response as well as a promise that returns the final aggregated response. |
+|  [countTokens(request, singleRequestOptions)](./vertexai.generativemodel.md#generativemodelcounttokens) |  | Counts the tokens in the provided request. |
+|  [generateContent(request, singleRequestOptions)](./vertexai.generativemodel.md#generativemodelgeneratecontent) |  | Makes a single non-streaming call to the model and returns an object containing a single <code>[GenerateContentResponse](./vertexai.generatecontentresponse.md#generatecontentresponse_interface)</code>. |
+|  [generateContentStream(request, singleRequestOptions)](./vertexai.generativemodel.md#generativemodelgeneratecontentstream) |  | Makes a single streaming call to the model and returns an object containing an iterable stream that iterates over all chunks in the streaming response as well as a promise that returns the final aggregated response. |
 |  [startChat(startChatParams)](./vertexai.generativemodel.md#generativemodelstartchat) |  | Gets a new <code>[ChatSession](./vertexai.chatsession.md#chatsession_class)</code> instance which can be used for multi-turn chats. |
 
 ## GenerativeModel.(constructor)
@@ -118,7 +118,7 @@ Counts the tokens in the provided request.
 <b>Signature:</b>
 
 ```typescript
-countTokens(request: CountTokensRequest | string | Array<string | Part>): Promise<CountTokensResponse>;
+countTokens(request: CountTokensRequest | string | Array<string | Part>, singleRequestOptions?: SingleRequestOptions): Promise<CountTokensResponse>;
 ```
 
 #### Parameters
@@ -126,6 +126,7 @@ countTokens(request: CountTokensRequest | string | Array<string | Part>): Promis
 |  Parameter | Type | Description |
 |  --- | --- | --- |
 |  request | [CountTokensRequest](./vertexai.counttokensrequest.md#counttokensrequest_interface) \| string \| Array&lt;string \| [Part](./vertexai.md#part)<!-- -->&gt; |  |
+|  singleRequestOptions | [SingleRequestOptions](./vertexai.singlerequestoptions.md#singlerequestoptions_interface) |  |
 
 <b>Returns:</b>
 
@@ -138,7 +139,7 @@ Makes a single non-streaming call to the model and returns an object containing
 <b>Signature:</b>
 
 ```typescript
-generateContent(request: GenerateContentRequest | string | Array<string | Part>): Promise<GenerateContentResult>;
+generateContent(request: GenerateContentRequest | string | Array<string | Part>, singleRequestOptions?: SingleRequestOptions): Promise<GenerateContentResult>;
 ```
 
 #### Parameters
@@ -146,6 +147,7 @@ generateContent(request: GenerateContentRequest | string | Array<string | Part>)
 |  Parameter | Type | Description |
 |  --- | --- | --- |
 |  request | [GenerateContentRequest](./vertexai.generatecontentrequest.md#generatecontentrequest_interface) \| string \| Array&lt;string \| [Part](./vertexai.md#part)<!-- -->&gt; |  |
+|  singleRequestOptions | [SingleRequestOptions](./vertexai.singlerequestoptions.md#singlerequestoptions_interface) |  |
 
 <b>Returns:</b>
 
@@ -158,7 +160,7 @@ Makes a single streaming call to the model and returns an object containing an i
 <b>Signature:</b>
 
 ```typescript
-generateContentStream(request: GenerateContentRequest | string | Array<string | Part>): Promise<GenerateContentStreamResult>;
+generateContentStream(request: GenerateContentRequest | string | Array<string | Part>, singleRequestOptions?: SingleRequestOptions): Promise<GenerateContentStreamResult>;
 ```
 
 #### Parameters
@@ -166,6 +168,7 @@ generateContentStream(request: GenerateContentRequest | string | Array<string |
 |  Parameter | Type | Description |
 |  --- | --- | --- |
 |  request | [GenerateContentRequest](./vertexai.generatecontentrequest.md#generatecontentrequest_interface) \| string \| Array&lt;string \| [Part](./vertexai.md#part)<!-- -->&gt; |  |
+|  singleRequestOptions | [SingleRequestOptions](./vertexai.singlerequestoptions.md#singlerequestoptions_interface) |  |
 
 <b>Returns:</b>
 
diff --git a/docs-devsite/vertexai.imagenmodel.md b/docs-devsite/vertexai.imagenmodel.md
index 63e15ff133a..d1f8885032c 100644
--- a/docs-devsite/vertexai.imagenmodel.md
+++ b/docs-devsite/vertexai.imagenmodel.md
@@ -42,7 +42,7 @@ export declare class ImagenModel extends VertexAIModel
 
 |  Method | Modifiers | Description |
 |  --- | --- | --- |
-|  [generateImages(prompt)](./vertexai.imagenmodel.md#imagenmodelgenerateimages) |  | <b><i>(Public Preview)</i></b> Generates images using the Imagen model and returns them as base64-encoded strings. |
+|  [generateImages(prompt, singleRequestOptions)](./vertexai.imagenmodel.md#imagenmodelgenerateimages) |  | <b><i>(Public Preview)</i></b> Generates images using the Imagen model and returns them as base64-encoded strings. |
 
 ## ImagenModel.(constructor)
 
@@ -118,7 +118,7 @@ If the prompt was not blocked, but one or more of the generated images were filt
 <b>Signature:</b>
 
 ```typescript
-generateImages(prompt: string): Promise<ImagenGenerationResponse<ImagenInlineImage>>;
+generateImages(prompt: string, singleRequestOptions?: SingleRequestOptions): Promise<ImagenGenerationResponse<ImagenInlineImage>>;
 ```
 
 #### Parameters
@@ -126,6 +126,7 @@ generateImages(prompt: string): Promise<ImagenGenerationResponse<ImagenInlineIma
 |  Parameter | Type | Description |
 |  --- | --- | --- |
 |  prompt | string | A text prompt describing the image(s) to generate. |
+|  singleRequestOptions | [SingleRequestOptions](./vertexai.singlerequestoptions.md#singlerequestoptions_interface) |  |
 
 <b>Returns:</b>
 
diff --git a/docs-devsite/vertexai.md b/docs-devsite/vertexai.md
index fca51b42f4f..cc411582bb4 100644
--- a/docs-devsite/vertexai.md
+++ b/docs-devsite/vertexai.md
@@ -111,6 +111,7 @@ The Vertex AI in Firebase Web SDK.
 |  [SchemaRequest](./vertexai.schemarequest.md#schemarequest_interface) | Final format for <code>[Schema](./vertexai.schema.md#schema_class)</code> params passed to backend requests. |
 |  [SchemaShared](./vertexai.schemashared.md#schemashared_interface) | Basic <code>[Schema](./vertexai.schema.md#schema_class)</code> properties shared across several Schema-related types. |
 |  [Segment](./vertexai.segment.md#segment_interface) |  |
+|  [SingleRequestOptions](./vertexai.singlerequestoptions.md#singlerequestoptions_interface) | Options that can be provided per-request. Extends the base [RequestOptions](./vertexai.requestoptions.md#requestoptions_interface) (like <code>timeout</code> and <code>baseUrl</code>) with request-specific controls like cancellation via <code>AbortSignal</code>.<!-- -->Options specified here will override any default [RequestOptions](./vertexai.requestoptions.md#requestoptions_interface) configured on a model (e.g. [GenerativeModel](./vertexai.generativemodel.md#generativemodel_class)<!-- -->). |
 |  [StartChatParams](./vertexai.startchatparams.md#startchatparams_interface) | Params for [GenerativeModel.startChat()](./vertexai.generativemodel.md#generativemodelstartchat)<!-- -->. |
 |  [TextPart](./vertexai.textpart.md#textpart_interface) | Content part interface if the part represents a text string. |
 |  [ToolConfig](./vertexai.toolconfig.md#toolconfig_interface) | Tool config. This config is shared for all tools provided in the request. |
diff --git a/docs-devsite/vertexai.singlerequestoptions.md b/docs-devsite/vertexai.singlerequestoptions.md
new file mode 100644
index 00000000000..af5d46059fd
--- /dev/null
+++ b/docs-devsite/vertexai.singlerequestoptions.md
@@ -0,0 +1,57 @@
+Project: /docs/reference/js/_project.yaml
+Book: /docs/reference/_book.yaml
+page_type: reference
+
+{% comment %}
+DO NOT EDIT THIS FILE!
+This is generated by the JS SDK team, and any local changes will be
+overwritten. Changes should be made in the source code at
+https://github.com/firebase/firebase-js-sdk
+{% endcomment %}
+
+# SingleRequestOptions interface
+Options that can be provided per-request. Extends the base [RequestOptions](./vertexai.requestoptions.md#requestoptions_interface) (like `timeout` and `baseUrl`<!-- -->) with request-specific controls like cancellation via `AbortSignal`<!-- -->.
+
+Options specified here will override any default [RequestOptions](./vertexai.requestoptions.md#requestoptions_interface) configured on a model (e.g. [GenerativeModel](./vertexai.generativemodel.md#generativemodel_class)<!-- -->).
+
+<b>Signature:</b>
+
+```typescript
+export interface SingleRequestOptions extends RequestOptions 
+```
+<b>Extends:</b> [RequestOptions](./vertexai.requestoptions.md#requestoptions_interface)
+
+## Properties
+
+|  Property | Type | Description |
+|  --- | --- | --- |
+|  [signal](./vertexai.singlerequestoptions.md#singlerequestoptionssignal) | AbortSignal | An <code>AbortSignal</code> instance that allows cancelling ongoing requests (like <code>generateContent</code> or <code>generateImages</code>).<!-- -->If provided, calling <code>abort()</code> on the corresponding <code>AbortController</code> will attempt to cancel the underlying HTTP request. An <code>AbortError</code> will be thrown if cancellation is successful.<!-- -->Note that this will not cancel the request in the backend, so billing will still be applied despite cancellation. |
+
+## SingleRequestOptions.signal
+
+An `AbortSignal` instance that allows cancelling ongoing requests (like `generateContent` or `generateImages`<!-- -->).
+
+If provided, calling `abort()` on the corresponding `AbortController` will attempt to cancel the underlying HTTP request. An `AbortError` will be thrown if cancellation is successful.
+
+Note that this will not cancel the request in the backend, so billing will still be applied despite cancellation.
+
+<b>Signature:</b>
+
+```typescript
+signal?: AbortSignal;
+```
+
+### Example
+
+
+```javascript
+const controller = new AbortController();
+const model = getGenerativeModel({
+  // ...
+});
+
+// To cancel request:
+controller.abort();
+
+```
+
diff --git a/packages/vertexai/src/methods/chat-session.ts b/packages/vertexai/src/methods/chat-session.ts
index dd22b29a7c8..42a741ef861 100644
--- a/packages/vertexai/src/methods/chat-session.ts
+++ b/packages/vertexai/src/methods/chat-session.ts
@@ -22,6 +22,7 @@ import {
   GenerateContentStreamResult,
   Part,
   RequestOptions,
+  SingleRequestOptions,
   StartChatParams
 } from '../types';
 import { formatNewContent } from '../requests/request-helpers';
@@ -75,7 +76,8 @@ export class ChatSession {
    * <code>{@link GenerateContentResult}</code>
    */
   async sendMessage(
-    request: string | Array<string | Part>
+    request: string | Array<string | Part>,
+    singleRequestOptions?: SingleRequestOptions
   ): Promise<GenerateContentResult> {
     await this._sendPromise;
     const newContent = formatNewContent(request);
@@ -95,7 +97,11 @@ export class ChatSession {
           this._apiSettings,
           this.model,
           generateContentRequest,
-          this.requestOptions
+          // Merge requestOptions
+          {
+            ...singleRequestOptions,
+            ...this.requestOptions
+          }
         )
       )
       .then(result => {
@@ -130,7 +136,8 @@ export class ChatSession {
    * and a response promise.
    */
   async sendMessageStream(
-    request: string | Array<string | Part>
+    request: string | Array<string | Part>,
+    singleRequestOptions?: SingleRequestOptions
   ): Promise<GenerateContentStreamResult> {
     await this._sendPromise;
     const newContent = formatNewContent(request);
@@ -146,7 +153,11 @@ export class ChatSession {
       this._apiSettings,
       this.model,
       generateContentRequest,
-      this.requestOptions
+      // Merge requestOptions
+      {
+        ...singleRequestOptions,
+        ...this.requestOptions
+      }
     );
 
     // Add onto the chain.
diff --git a/packages/vertexai/src/methods/count-tokens.ts b/packages/vertexai/src/methods/count-tokens.ts
index c9d43a5b6fd..b89dab44e64 100644
--- a/packages/vertexai/src/methods/count-tokens.ts
+++ b/packages/vertexai/src/methods/count-tokens.ts
@@ -18,7 +18,7 @@
 import {
   CountTokensRequest,
   CountTokensResponse,
-  RequestOptions
+  SingleRequestOptions
 } from '../types';
 import { Task, makeRequest } from '../requests/request';
 import { ApiSettings } from '../types/internal';
@@ -27,7 +27,7 @@ export async function countTokens(
   apiSettings: ApiSettings,
   model: string,
   params: CountTokensRequest,
-  requestOptions?: RequestOptions
+  singleRequestOptions?: SingleRequestOptions
 ): Promise<CountTokensResponse> {
   const response = await makeRequest(
     model,
@@ -35,7 +35,7 @@ export async function countTokens(
     apiSettings,
     false,
     JSON.stringify(params),
-    requestOptions
+    singleRequestOptions
   );
   return response.json();
 }
diff --git a/packages/vertexai/src/methods/generate-content.ts b/packages/vertexai/src/methods/generate-content.ts
index 0944b38016a..c67c36ac1d2 100644
--- a/packages/vertexai/src/methods/generate-content.ts
+++ b/packages/vertexai/src/methods/generate-content.ts
@@ -20,7 +20,7 @@ import {
   GenerateContentResponse,
   GenerateContentResult,
   GenerateContentStreamResult,
-  RequestOptions
+  SingleRequestOptions
 } from '../types';
 import { Task, makeRequest } from '../requests/request';
 import { createEnhancedContentResponse } from '../requests/response-helpers';
@@ -31,7 +31,7 @@ export async function generateContentStream(
   apiSettings: ApiSettings,
   model: string,
   params: GenerateContentRequest,
-  requestOptions?: RequestOptions
+  singleRequestOptions?: SingleRequestOptions
 ): Promise<GenerateContentStreamResult> {
   const response = await makeRequest(
     model,
@@ -39,7 +39,7 @@ export async function generateContentStream(
     apiSettings,
     /* stream */ true,
     JSON.stringify(params),
-    requestOptions
+    singleRequestOptions
   );
   return processStream(response);
 }
@@ -48,7 +48,7 @@ export async function generateContent(
   apiSettings: ApiSettings,
   model: string,
   params: GenerateContentRequest,
-  requestOptions?: RequestOptions
+  singleRequestOptions?: SingleRequestOptions
 ): Promise<GenerateContentResult> {
   const response = await makeRequest(
     model,
@@ -56,7 +56,7 @@ export async function generateContent(
     apiSettings,
     /* stream */ false,
     JSON.stringify(params),
-    requestOptions
+    singleRequestOptions
   );
   const responseJson: GenerateContentResponse = await response.json();
   const enhancedResponse = createEnhancedContentResponse(responseJson);
diff --git a/packages/vertexai/src/models/generative-model.ts b/packages/vertexai/src/models/generative-model.ts
index b4cf464f025..5991e9bd62c 100644
--- a/packages/vertexai/src/models/generative-model.ts
+++ b/packages/vertexai/src/models/generative-model.ts
@@ -29,11 +29,12 @@ import {
   GenerationConfig,
   ModelParams,
   Part,
-  RequestOptions,
   SafetySetting,
+  RequestOptions,
   StartChatParams,
   Tool,
-  ToolConfig
+  ToolConfig,
+  SingleRequestOptions
 } from '../types';
 import { ChatSession } from '../methods/chat-session';
 import { countTokens } from '../methods/count-tokens';
@@ -77,7 +78,8 @@ export class GenerativeModel extends VertexAIModel {
    * and returns an object containing a single <code>{@link GenerateContentResponse}</code>.
    */
   async generateContent(
-    request: GenerateContentRequest | string | Array<string | Part>
+    request: GenerateContentRequest | string | Array<string | Part>,
+    singleRequestOptions?: SingleRequestOptions
   ): Promise<GenerateContentResult> {
     const formattedParams = formatGenerateContentInput(request);
     return generateContent(
@@ -91,7 +93,11 @@ export class GenerativeModel extends VertexAIModel {
         systemInstruction: this.systemInstruction,
         ...formattedParams
       },
-      this.requestOptions
+      // Merge request options
+      {
+        ...singleRequestOptions,
+        ...this.requestOptions
+      }
     );
   }
 
@@ -102,7 +108,8 @@ export class GenerativeModel extends VertexAIModel {
    * a promise that returns the final aggregated response.
    */
   async generateContentStream(
-    request: GenerateContentRequest | string | Array<string | Part>
+    request: GenerateContentRequest | string | Array<string | Part>,
+    singleRequestOptions?: SingleRequestOptions
   ): Promise<GenerateContentStreamResult> {
     const formattedParams = formatGenerateContentInput(request);
     return generateContentStream(
@@ -116,7 +123,11 @@ export class GenerativeModel extends VertexAIModel {
         systemInstruction: this.systemInstruction,
         ...formattedParams
       },
-      this.requestOptions
+      // Merge request options
+      {
+        ...singleRequestOptions,
+        ...this.requestOptions
+      }
     );
   }
 
@@ -142,9 +153,19 @@ export class GenerativeModel extends VertexAIModel {
    * Counts the tokens in the provided request.
    */
   async countTokens(
-    request: CountTokensRequest | string | Array<string | Part>
+    request: CountTokensRequest | string | Array<string | Part>,
+    singleRequestOptions?: SingleRequestOptions
   ): Promise<CountTokensResponse> {
     const formattedParams = formatGenerateContentInput(request);
-    return countTokens(this._apiSettings, this.model, formattedParams);
+    return countTokens(
+      this._apiSettings,
+      this.model,
+      formattedParams,
+      // Merge request options
+      {
+        ...singleRequestOptions,
+        ...this.requestOptions
+      }
+    );
   }
 }
diff --git a/packages/vertexai/src/models/imagen-model.ts b/packages/vertexai/src/models/imagen-model.ts
index 89c740852a3..7e7de956aa9 100644
--- a/packages/vertexai/src/models/imagen-model.ts
+++ b/packages/vertexai/src/models/imagen-model.ts
@@ -26,7 +26,8 @@ import {
   RequestOptions,
   ImagenModelParams,
   ImagenGenerationResponse,
-  ImagenSafetySettings
+  ImagenSafetySettings,
+  SingleRequestOptions
 } from '../types';
 import { VertexAIModel } from './vertexai-model';
 
@@ -102,7 +103,8 @@ export class ImagenModel extends VertexAIModel {
    * @beta
    */
   async generateImages(
-    prompt: string
+    prompt: string,
+    singleRequestOptions?: SingleRequestOptions
   ): Promise<ImagenGenerationResponse<ImagenInlineImage>> {
     const body = createPredictRequestBody(prompt, {
       ...this.generationConfig,
@@ -114,7 +116,11 @@ export class ImagenModel extends VertexAIModel {
       this._apiSettings,
       /* stream */ false,
       JSON.stringify(body),
-      this.requestOptions
+      // Merge request options
+      {
+        ...singleRequestOptions,
+        ...this.requestOptions
+      }
     );
     return handlePredictResponse<ImagenInlineImage>(response);
   }
@@ -140,7 +146,8 @@ export class ImagenModel extends VertexAIModel {
    */
   async generateImagesGCS(
     prompt: string,
-    gcsURI: string
+    gcsURI: string,
+    singleRequestOptions?: SingleRequestOptions
   ): Promise<ImagenGenerationResponse<ImagenGCSImage>> {
     const body = createPredictRequestBody(prompt, {
       gcsURI,
@@ -153,7 +160,10 @@ export class ImagenModel extends VertexAIModel {
       this._apiSettings,
       /* stream */ false,
       JSON.stringify(body),
-      this.requestOptions
+      {
+        ...singleRequestOptions,
+        ...this.requestOptions
+      }
     );
     return handlePredictResponse<ImagenGCSImage>(response);
   }
diff --git a/packages/vertexai/src/requests/request.ts b/packages/vertexai/src/requests/request.ts
index 47e4c6ab446..002930cfbc8 100644
--- a/packages/vertexai/src/requests/request.ts
+++ b/packages/vertexai/src/requests/request.ts
@@ -15,7 +15,11 @@
  * limitations under the License.
  */
 
-import { ErrorDetails, RequestOptions, VertexAIErrorCode } from '../types';
+import {
+  ErrorDetails,
+  SingleRequestOptions,
+  VertexAIErrorCode
+} from '../types';
 import { VertexAIError } from '../errors';
 import { ApiSettings } from '../types/internal';
 import {
@@ -27,6 +31,9 @@ import {
 } from '../constants';
 import { logger } from '../logger';
 
+const TIMEOUT_EXPIRED_MESSAGE = 'Timeout has expired.';
+const ABORT_ERROR_NAME = 'AbortError';
+
 export enum Task {
   GENERATE_CONTENT = 'generateContent',
   STREAM_GENERATE_CONTENT = 'streamGenerateContent',
@@ -40,7 +47,7 @@ export class RequestUrl {
     public task: Task,
     public apiSettings: ApiSettings,
     public stream: boolean,
-    public requestOptions?: RequestOptions
+    public requestOptions?: SingleRequestOptions
   ) {}
   toString(): string {
     // TODO: allow user-set option if that feature becomes available
@@ -115,9 +122,15 @@ export async function constructRequest(
   apiSettings: ApiSettings,
   stream: boolean,
   body: string,
-  requestOptions?: RequestOptions
+  singleRequestOptions?: SingleRequestOptions
 ): Promise<{ url: string; fetchOptions: RequestInit }> {
-  const url = new RequestUrl(model, task, apiSettings, stream, requestOptions);
+  const url = new RequestUrl(
+    model,
+    task,
+    apiSettings,
+    stream,
+    singleRequestOptions
+  );
   return {
     url: url.toString(),
     fetchOptions: {
@@ -134,11 +147,49 @@ export async function makeRequest(
   apiSettings: ApiSettings,
   stream: boolean,
   body: string,
-  requestOptions?: RequestOptions
+  singleRequestOptions?: SingleRequestOptions
 ): Promise<Response> {
-  const url = new RequestUrl(model, task, apiSettings, stream, requestOptions);
+  const url = new RequestUrl(
+    model,
+    task,
+    apiSettings,
+    stream,
+    singleRequestOptions
+  );
   let response;
-  let fetchTimeoutId: string | number | NodeJS.Timeout | undefined;
+
+  const externalSignal = singleRequestOptions?.signal;
+  const timeoutMillis =
+    singleRequestOptions?.timeout != null && singleRequestOptions.timeout >= 0
+      ? singleRequestOptions.timeout
+      : DEFAULT_FETCH_TIMEOUT_MS;
+  const internalAbortController = new AbortController();
+  const fetchTimeoutId = setTimeout(() => {
+    internalAbortController.abort(TIMEOUT_EXPIRED_MESSAGE);
+    logger.debug(
+      `Aborting request to ${url} due to timeout (${timeoutMillis}ms)`
+    );
+  }, timeoutMillis);
+
+  if (externalSignal) {
+    if (externalSignal.aborted) {
+      clearTimeout(fetchTimeoutId);
+      throw new DOMException(
+        externalSignal.reason ?? 'Aborted externally before fetch',
+        ABORT_ERROR_NAME
+      );
+    }
+
+    const externalAbortListener = (): void => {
+      logger.debug(`Aborting request to ${url} due to external abort signal.`);
+      internalAbortController.abort(externalSignal.reason);
+    };
+
+    externalSignal.addEventListener('abort', externalAbortListener, {
+      once: true
+    });
+  }
+
   try {
     const request = await constructRequest(
       model,
@@ -146,16 +197,9 @@ export async function makeRequest(
       apiSettings,
       stream,
       body,
-      requestOptions
+      singleRequestOptions
     );
-    // Timeout is 180s by default
-    const timeoutMillis =
-      requestOptions?.timeout != null && requestOptions.timeout >= 0
-        ? requestOptions.timeout
-        : DEFAULT_FETCH_TIMEOUT_MS;
-    const abortController = new AbortController();
-    fetchTimeoutId = setTimeout(() => abortController.abort(), timeoutMillis);
-    request.fetchOptions.signal = abortController.signal;
+    request.fetchOptions.signal = internalAbortController.signal;
 
     response = await fetch(request.url, request.fetchOptions);
     if (!response.ok) {
diff --git a/packages/vertexai/src/types/requests.ts b/packages/vertexai/src/types/requests.ts
index 5058b457365..f1fb7536d12 100644
--- a/packages/vertexai/src/types/requests.ts
+++ b/packages/vertexai/src/types/requests.ts
@@ -143,6 +143,44 @@ export interface RequestOptions {
   baseUrl?: string;
 }
 
+/**
+ * Options that can be provided per-request.
+ * Extends the base {@link RequestOptions} (like `timeout` and `baseUrl`)
+ * with request-specific controls like cancellation via `AbortSignal`.
+ *
+ * Options specified here will override any default {@link RequestOptions}
+ * configured on a model (e.g. {@link GenerativeModel}).
+ *
+ * @public
+ */
+export interface SingleRequestOptions extends RequestOptions {
+  /**
+   * An `AbortSignal` instance that allows cancelling ongoing requests (like `generateContent` or
+   * `generateImages`).
+   *
+   * If provided, calling `abort()` on the corresponding `AbortController`
+   * will attempt to cancel the underlying HTTP request. An `AbortError` will be thrown
+   * if cancellation is successful.
+   *
+   * Note that this will not cancel the request in the backend, so billing will
+   * still be applied despite cancellation.
+   *
+   * @example
+   * ```javascript
+   * const controller = new AbortController();
+   * const model = getGenerativeModel({
+   *   // ...
+   * });
+   *
+   * // To cancel request:
+   * controller.abort();
+   * ```
+   * @see https://developer.mozilla.org/en-US/docs/Web/API/AbortSignal
+   */
+
+  signal?: AbortSignal;
+}
+
 /**
  * Defines a tool that model can call to access external knowledge.
  * @public