diff --git a/src/renderer/src/aiCore/utils/__tests__/options.test.ts b/src/renderer/src/aiCore/utils/__tests__/options.test.ts
index 616269a054..d3622e8696 100644
--- a/src/renderer/src/aiCore/utils/__tests__/options.test.ts
+++ b/src/renderer/src/aiCore/utils/__tests__/options.test.ts
@@ -1123,6 +1123,46 @@ describe('options utils', () => {
         })
       })
 
+      it.each([
+        { providerId: 'newapi', providerName: 'NewAPI' },
+        { providerId: 'aihubmix', providerName: 'AiHubMix' }
+      ])(
+        'should route Gemini models to google providerOptions through $providerName',
+        async ({ providerId, providerName }) => {
+          const { getCustomParameters } = await import('../reasoning')
+          vi.mocked(getCustomParameters).mockReturnValue({
+            generationConfig: { responseModalities: ['IMAGE', 'TEXT'] },
+            imageConfig: { aspectRatio: '3:4', imageSize: '4K' }
+          })
+
+          const provider: Provider = {
+            id: providerId,
+            name: providerName,
+            type: 'openai',
+            models: [] as Model[]
+          } as Provider
+
+          const geminiModel: Model = {
+            id: 'gemini-3.1-flash-image-preview',
+            name: 'Gemini 3.1 Flash Image Preview',
+            provider: providerId
+          } as Model
+
+          const result = buildProviderOptions(mockAssistant, geminiModel, provider, {
+            enableReasoning: false,
+            enableWebSearch: false,
+            enableGenerateImage: true
+          })
+
+          expect(result.providerOptions).toHaveProperty('google')
+          expect(result.providerOptions).not.toHaveProperty(providerId)
+          expect(result.providerOptions.google).toMatchObject({
+            generationConfig: { responseModalities: ['IMAGE', 'TEXT'] },
+            imageConfig: { aspectRatio: '3:4', imageSize: '4K' }
+          })
+        }
+      )
+
       // Note: For proxy providers like aihubmix/newapi, users should write AI SDK provider ID (google/anthropic)
       // instead of the Cherry Studio provider ID for custom parameters to work correctly
 
diff --git a/src/renderer/src/aiCore/utils/options.ts b/src/renderer/src/aiCore/utils/options.ts
index 1f2ace67db..0bee602478 100644
--- a/src/renderer/src/aiCore/utils/options.ts
+++ b/src/renderer/src/aiCore/utils/options.ts
@@ -205,6 +205,8 @@ export function buildProviderOptions(
     case SystemProviderIds.ollama:
       providerSpecificOptions = buildOllamaProviderOptions(assistant, model, capabilities)
       break
+    case 'newapi':
+    case 'aihubmix':
     case SystemProviderIds.gateway:
       providerSpecificOptions = buildAIGatewayOptions(assistant, model, capabilities, serviceTier, textVerbosity)
       break
diff --git a/src/renderer/src/config/models/vision.ts b/src/renderer/src/config/models/vision.ts
index 0faf8ba5ab..10c832cfee 100644
--- a/src/renderer/src/config/models/vision.ts
+++ b/src/renderer/src/config/models/vision.ts
@@ -119,7 +119,7 @@ const IMAGE_ENHANCEMENT_MODELS = [
   'gpt-image-1',
   'gemini-2.5-flash-image(?:-[\\w-]+)?',
   'gemini-2.0-flash-preview-image-generation',
-  'gemini-3(?:\\.\\d+)?-pro-image(?:-[\\w-]+)?'
+  'gemini-3(?:\\.\\d+)?-(?:flash|pro)-image(?:-[\\w-]+)?'
 ]
 
 const IMAGE_ENHANCEMENT_MODELS_REGEX = new RegExp(IMAGE_ENHANCEMENT_MODELS.join('|'), 'i')
@@ -129,7 +129,7 @@ const DEDICATED_IMAGE_MODEL_REGEX = new RegExp(DEDICATED_IMAGE_MODELS.join('|'),
 // Models that should auto-enable image generation button when selected
 const AUTO_ENABLE_IMAGE_MODELS = [
   'gemini-2.5-flash-image(?:-[\\w-]+)?',
-  'gemini-3(?:\\.\\d+)?-pro-image(?:-[\\w-]+)?',
+  'gemini-3(?:\\.\\d+)?-(?:flash|pro)-image(?:-[\\w-]+)?',
   ...DEDICATED_IMAGE_MODELS
 ]
 
@@ -147,7 +147,7 @@ const OPENAI_TOOL_USE_IMAGE_GENERATION_MODELS = [
 
 const OPENAI_IMAGE_GENERATION_MODELS = [...OPENAI_TOOL_USE_IMAGE_GENERATION_MODELS, 'gpt-image-1']
 
-const MODERN_IMAGE_MODELS = ['gemini-3(?:\\.\\d+)?-pro-image(?:-[\\w-]+)?']
+const MODERN_IMAGE_MODELS = ['gemini-3(?:\\.\\d+)?-(?:flash|pro)-image(?:-[\\w-]+)?']
 
 const GENERATE_IMAGE_MODELS = [
   'gemini-2.0-flash-exp(?:-[\\w-]+)?',