diff --git a/build/build_number.txt b/build/build_number.txt
index dc37bbd..9870ccc 100644
--- a/build/build_number.txt
+++ b/build/build_number.txt
@@ -1 +1 @@
-184
+187
diff --git a/tests/unit/engine_google.test.ts b/tests/unit/engine_google.test.ts
new file mode 100644
index 0000000..076b712
--- /dev/null
+++ b/tests/unit/engine_google.test.ts
@@ -0,0 +1,104 @@
+
+import { vi, beforeEach, expect, test } from 'vitest'
+import { store } from '../../src/services/store'
+import defaults from '../../defaults/settings.json'
+import Message from '../../src/models/message'
+import Google from '../../src/services/google'
+import { EnhancedGenerateContentResponse } from '@google/generative-ai'
+
+vi.mock('@google/generative-ai', async() => {
+  return {
+    GoogleGenerativeAI: vi.fn((apiKey) => {
+      return {
+        apiKey: apiKey,
+        getGenerativeModel: vi.fn(() => {
+          return {
+            startChat: vi.fn(() => {
+              return {
+                sendMessage: vi.fn(() => {
+                  return {
+                    response: {
+                      text: () => 'response'
+                    }
+                  }
+                }),
+                sendMessageStream: vi.fn(() => {
+                  return {
+                    stream: vi.fn()
+                  }
+                })
+              }
+            })
+          }
+        }),
+      }
+    })
+  }
+})
+
+beforeEach(() => {
+  store.config = defaults
+  store.config.engines.google.apiKey = '123'
+})
+
+test('Google Basic', async () => {
+  const google = new Google(store.config)
+  expect(google.getName()).toBe('google')
+  expect(await google.getModels()).toStrictEqual([
+    { id: 'models/gemini-1.5-pro-latest', name: 'Gemini 1.5 Pro' },
+    { id: 'gemini-1.5-flash-latest', name: 'Gemini  1.5 Flash' },
+    { id: 'models/gemini-pro', name: 'Gemini 1.0 Pro' },
+])
+  //expect(_Google.default.prototype.models.list).toHaveBeenCalled()
+  expect(google.isVisionModel('models/gemini-pro')).toBe(false)
+  expect(google.isVisionModel('gemini-1.5-flash-latest')).toBe(true)
+  expect(google.isVisionModel('models/gemini-1.5-pro-latest')).toBe(true)
+  expect(google.getRountingModel()).toBeNull()
+})
+
+test('Google completion', async () => {
+  const google = new Google(store.config)
+  const response = await google.complete([
+    new Message('system', 'instruction'),
+    new Message('user', 'prompt'),
+  ], null)
+  expect(response).toStrictEqual({
+    type: 'text',
+    content: 'response'
+  })
+})
+
+test('Google stream', async () => {
+  const google = new Google(store.config)
+  const response = await google.stream([
+    new Message('system', 'instruction'),
+    new Message('user', 'prompt'),
+  ], null)
+  //expect(response.controller).toBeDefined()
+  await google.stop(response)
+  //expect(response.controller.abort).toHaveBeenCalled()
+})
+
+test('Google streamChunkToLlmChunk Text', async () => {
+  const google = new Google(store.config)
+  const streamChunk: EnhancedGenerateContentResponse = {
+    candidates: [ {
+      index: 0,
+      content: { role: 'model', parts: [ { text: 'response' } ] },
+      //finishReason: FinishReason.STOP,
+    } ],
+    text: vi.fn(() => 'response'),
+    functionCalls: vi.fn(() => []),
+    functionCall: null,
+  }
+  const llmChunk1 = await google.streamChunkToLlmChunk(streamChunk, null)
+  expect(streamChunk.text).toHaveBeenCalled()
+  expect(streamChunk.functionCalls).toHaveBeenCalled()
+  expect(llmChunk1).toStrictEqual({ text: 'response', done: false })
+  streamChunk.candidates[0].finishReason = 'STOP'
+  streamChunk.text = vi.fn(() => '')
+  const llmChunk2 = await google.streamChunkToLlmChunk(streamChunk, null)
+  expect(streamChunk.text).toHaveBeenCalled()
+  expect(streamChunk.functionCalls).toHaveBeenCalled()
+  expect(llmChunk2).toStrictEqual({ text: '', done: true })
+})