test(google_genai): move unified_google_tests to gemini-3.5-flash-lite (#42520)

gemini-2.5-flash-lite retires on 2026-10-20 per Google's Vertex model
lifecycle table, which would turn the google_generate_content_endpoint_testing
CircleCI job red on every PR from that day. gemini-3.5-flash-lite is the newest
flash-lite on both AI Studio and Vertex, with a retirement of 2027-07-21 or
later.

Vertex serves gemini-3.5-flash-lite only on the global, us, and eu endpoints,
never on a single region, so the direct SDK Vertex tests now pass
vertex_location global like the suite's proxy config already does.

The 429 backoff test derives its mocked generateContent URL from litellm's own
Gemini URL builder, since litellm routes Gemini 3 and newer models to v1alpha
instead of v1beta.

Co-authored-by: mateo-berri <277851410+mateo-berri@users.noreply.github.com>

tests/unified_google_tests/conftest.py

@@ -102,7 +102,7 @@ def google_genai_proxy_url() -> Iterator[str]:
         credentials_file = os.environ.get("GOOGLE_APPLICATION_CREDENTIALS", "")
         if not (credentials_file and os.path.isfile(credentials_file)):
             vertex_credentials_path = load_vertex_ai_credentials(
-                model="vertex_ai/gemini-2.5-flash-lite"
+                model="vertex_ai/gemini-3.5-flash-lite"
             )
             if vertex_credentials_path:
                 temp_credentials_path = vertex_credentials_path

tests/unified_google_tests/google_genai_proxy_test_config.yaml

@@ -1,12 +1,12 @@
 model_list:
-  - model_name: gemini-2.5-flash-lite
+  - model_name: gemini-3.5-flash-lite
     litellm_params:
-      model: gemini/gemini-2.5-flash-lite
+      model: gemini/gemini-3.5-flash-lite
       api_key: os.environ/GEMINI_API_KEY
 
-  - model_name: vertex-gemini-2.5-flash-lite
+  - model_name: vertex-gemini-3.5-flash-lite
     litellm_params:
-      model: vertex_ai/gemini-2.5-flash-lite
+      model: vertex_ai/gemini-3.5-flash-lite
       vertex_location: global
 
 router_settings:

tests/unified_google_tests/test_google_ai_studio.py

@@ -13,12 +13,12 @@ class TestGoogleGenAIStudio(BaseGoogleGenAITest, BaseGoogleGenAIProxySDKTest):
     @property
     def model_config(self):
         return {
-            "model": "gemini/gemini-2.5-flash-lite",
+            "model": "gemini/gemini-3.5-flash-lite",
         }
 
     @property
     def proxy_model_name(self) -> str:
-        return "gemini-2.5-flash-lite"
+        return "gemini-3.5-flash-lite"
 
 
 @pytest.mark.asyncio
@@ -94,7 +94,7 @@ async def test_mock_stream_generate_content_with_tools():
             "\n--- Testing async agenerate_content_stream with function call parsing ---"
         )
         response = await litellm.google_genai.agenerate_content_stream(
-            model="gemini/gemini-2.5-flash-lite",
+            model="gemini/gemini-3.5-flash-lite",
             contents=contents,
             tools=[
                 {
@@ -343,7 +343,7 @@ async def test_validate_post_request_parameters():
 
         # Make the API call
         response = await litellm.google_genai.agenerate_content_stream(
-            model="gemini/gemini-2.5-flash-lite", contents=contents, tools=tools
+            model="gemini/gemini-3.5-flash-lite", contents=contents, tools=tools
         )
 
         # Consume the response to ensure the request is made
@@ -387,11 +387,11 @@ async def test_validate_post_request_parameters():
 
         # Validate model field
         assert "model" in request_data, "Expected 'model' field in request data"
-        # Model might be transformed, but should contain gemini-2.5-flash-lite
+        # Model might be transformed, but should contain gemini-3.5-flash-lite
         model_value = request_data["model"]
         assert (
-            "gemini-2.5-flash-lite" in model_value
-        ), f"Expected model to contain 'gemini-2.5-flash-lite', got: {model_value}"
+            "gemini-3.5-flash-lite" in model_value
+        ), f"Expected model to contain 'gemini-3.5-flash-lite', got: {model_value}"
         print(f"✅ Model validation passed: {model_value}")
 
         # Validate contents field

tests/unified_google_tests/test_google_genai_proxy_test_config.py

@@ -12,14 +12,13 @@ from typing_extensions import ReadOnly, TypedDict
 import litellm
 from litellm import Router
 from litellm.constants import INITIAL_RETRY_DELAY, MAX_RETRY_DELAY
-from litellm.llms.vertex_ai.common_utils import get_vertex_base_url
+from litellm.llms.vertex_ai.common_utils import _get_gemini_url, get_vertex_base_url
 from litellm.llms.vertex_ai.vertex_llm_base import VertexBase
 
 CONFIG_PATH: Final = Path(__file__).parent / "google_genai_proxy_test_config.yaml"
-GEMINI_DEPLOYMENT: Final = "gemini-2.5-flash-lite"
-VERTEX_DEPLOYMENT: Final = "vertex-gemini-2.5-flash-lite"
-GEMINI_HOST: Final = "generativelanguage.googleapis.com"
-GEMINI_GENERATE_CONTENT_PATH: Final = "/v1beta/models/gemini-2.5-flash-lite:generateContent"
+GEMINI_DEPLOYMENT: Final = "gemini-3.5-flash-lite"
+VERTEX_DEPLOYMENT: Final = "vertex-gemini-3.5-flash-lite"
+GEMINI_GENERATE_CONTENT_URL: Final = _get_gemini_url(mode="chat", model=GEMINI_DEPLOYMENT, stream=False)[0]
 VERTEX_GLOBAL_BASE_URL: Final = "https://aiplatform.googleapis.com"
 RESOURCE_EXHAUSTED: Final = {
     "error": {"code": 429, "message": "Resource exhausted. Please try again later.", "status": "RESOURCE_EXHAUSTED"}
@@ -81,7 +80,7 @@ async def test_ci_proxy_config_rides_out_consecutive_429s_with_backoff(
 ) -> None:
     monkeypatch.setattr(litellm, "disable_aiohttp_transport", True)
     litellm.in_memory_llm_clients_cache.flush_cache()
-    route: Final = respx_mock.post(host=GEMINI_HOST, path=GEMINI_GENERATE_CONTENT_PATH).mock(
+    route: Final = respx_mock.post(GEMINI_GENERATE_CONTENT_URL).mock(
         side_effect=[httpx.Response(429, json=RESOURCE_EXHAUSTED)] * CONSECUTIVE_RATE_LIMITS
         + [httpx.Response(200, json=PONG)]
     )

tests/unified_google_tests/test_vertex_ai_native.py

@@ -8,9 +8,10 @@ class TestVertexAIGenerateContent(BaseGoogleGenAITest, BaseGoogleGenAIProxySDKTe
     @property
     def model_config(self):
         return {
-            "model": "vertex_ai/gemini-2.5-flash-lite",
+            "model": "vertex_ai/gemini-3.5-flash-lite",
+            "vertex_location": "global",
         }
 
     @property
     def proxy_model_name(self) -> str:
-        return "vertex-gemini-2.5-flash-lite"
+        return "vertex-gemini-3.5-flash-lite"