diff --git a/packages/ai/integration/constants.ts b/packages/ai/integration/constants.ts
index 99a65f31c54..3e84f360a03 100644
--- a/packages/ai/integration/constants.ts
+++ b/packages/ai/integration/constants.ts
@@ -57,7 +57,7 @@ const modelNames: readonly string[] = ['gemini-2.0-flash', 'gemini-2.5-flash'];
 // The Live API requires a different set of models, and they're different for each backend.
 const liveModelNames: Map<BackendType, string[]> = new Map([
   [BackendType.GOOGLE_AI, ['gemini-live-2.5-flash-preview']],
-  [BackendType.VERTEX_AI, ['gemini-2.0-flash-exp']]
+  [BackendType.VERTEX_AI, ['gemini-2.0-flash-live-preview-04-09']]
 ]);
 
 /**
diff --git a/packages/ai/integration/live.test.ts b/packages/ai/integration/live.test.ts
index 6b50fe65222..2a357951641 100644
--- a/packages/ai/integration/live.test.ts
+++ b/packages/ai/integration/live.test.ts
@@ -17,7 +17,6 @@
 
 import { expect } from 'chai';
 import {
-  BackendType,
   getLiveGenerativeModel,
   LiveGenerationConfig,
   LiveServerContent,
@@ -81,9 +80,6 @@ describe('Live', function () {
   };
 
   liveTestConfigs.forEach(testConfig => {
-    if (testConfig.ai.backend.backendType === BackendType.VERTEX_AI) {
-      return;
-    }
     describe(`${testConfig.toString()}`, () => {
       describe('Live', () => {
         it('should connect, send a message, receive a response, and close', async () => {