fix: preserve live thinking config

src/google/adk/models/google_llm.py

@@ -453,6 +453,10 @@ async def connect(self, llm_request: LlmRequest) -> BaseLlmConnection:
             ' backend. Please use Vertex AI backend.'
         )
     llm_request.live_connect_config.tools = llm_request.config.tools
+    if llm_request.config.thinking_config is not None:
+      llm_request.live_connect_config.thinking_config = (
+          llm_request.config.thinking_config
+      )
     logger.debug('Connecting to live with llm_request:%s', llm_request)
     logger.debug('Live connect config: %s', llm_request.live_connect_config)
     async with self._live_api_client.aio.live.connect(

tests/unittests/models/test_google_llm.py

@@ -40,6 +40,17 @@
 from google.genai.types import Part
 import pytest
 
+_GENERATIVE_LANGUAGE_API_BASE_URL = (
+    "https://generativelanguage.googleapis.com/v1alpha"
+)
+_GENERATIVE_LANGUAGE_API_ROOT = "https://generativelanguage.googleapis.com/"
+_GENERATIVE_LANGUAGE_MTLS_API_BASE_URL = (
+    "https://generativelanguage.mtls.googleapis.com/v1alpha"
+)
+_GENERATIVE_LANGUAGE_MTLS_API_ROOT = (
+    "https://generativelanguage.mtls.googleapis.com/"
+)
+
 
 class MockAsyncIterator:
   """Mock for async iterator."""
@@ -252,17 +263,27 @@ def test_client_version_header_with_agent_engine(monkeypatch):
   )
 
 
-def test_api_client_uses_api_version_from_google_base_url():
+@pytest.mark.parametrize(
+    ("base_url", "normalized_base_url"),
+    (
+        (_GENERATIVE_LANGUAGE_API_BASE_URL, _GENERATIVE_LANGUAGE_API_ROOT),
+        (
+            _GENERATIVE_LANGUAGE_MTLS_API_BASE_URL,
+            _GENERATIVE_LANGUAGE_MTLS_API_ROOT,
+        ),
+    ),
+)
+def test_api_client_uses_api_version_from_google_base_url(
+    base_url, normalized_base_url
+):
   model = Gemini(
       model="gemini-2.5-flash",
-      base_url="https://generativelanguage.googleapis.com/v1alpha",
+      base_url=base_url,
   )
 
   client = model.api_client
 
-  assert client._api_client._http_options.base_url == (
-      "https://generativelanguage.googleapis.com/"
-  )
+  assert client._api_client._http_options.base_url == normalized_base_url
   assert client._api_client._http_options.api_version == "v1alpha"
 
 
@@ -670,7 +691,7 @@ async def test_generate_content_async_patches_api_version(
 ):
   gemini_llm = Gemini(
       model="gemini-2.5-flash",
-      base_url="https://generativelanguage.googleapis.com/v1alpha",
+      base_url=_GENERATIVE_LANGUAGE_API_BASE_URL,
   )
   llm_request.config.http_options = types.HttpOptions(
       headers={"custom-header": "custom-value"}
@@ -718,7 +739,7 @@ def test_live_api_version_vertex_ai(gemini_llm):
 def test_live_api_version_uses_google_base_url_version():
   gemini_llm = Gemini(
       model="gemini-2.5-flash",
-      base_url="https://generativelanguage.googleapis.com/v1alpha",
+      base_url=_GENERATIVE_LANGUAGE_API_BASE_URL,
   )
 
   assert gemini_llm._live_api_version == "v1alpha"
@@ -732,16 +753,28 @@ def test_live_api_version_gemini_api(gemini_llm):
     assert gemini_llm._live_api_version == "v1alpha"
 
 
-def test_live_api_client_uses_api_version_from_google_base_url():
+@pytest.mark.parametrize(
+    ("base_url", "normalized_base_url"),
+    (
+        (_GENERATIVE_LANGUAGE_API_BASE_URL, _GENERATIVE_LANGUAGE_API_ROOT),
+        (
+            _GENERATIVE_LANGUAGE_MTLS_API_BASE_URL,
+            _GENERATIVE_LANGUAGE_MTLS_API_ROOT,
+        ),
+    ),
+)
+def test_live_api_client_uses_api_version_from_google_base_url(
+    base_url, normalized_base_url
+):
   gemini_llm = Gemini(
       model="gemini-2.5-flash",
-      base_url="https://generativelanguage.googleapis.com/v1alpha",
+      base_url=base_url,
   )
 
   client = gemini_llm._live_api_client
   http_options = client._api_client._http_options
 
-  assert http_options.base_url == "https://generativelanguage.googleapis.com/"
+  assert http_options.base_url == normalized_base_url
   assert http_options.api_version == "v1alpha"
 
 
@@ -852,6 +885,38 @@ async def __aexit__(self, *args):
         )
 
 
+@pytest.mark.asyncio
+async def test_connect_copies_thinking_config_to_live_config(
+    gemini_llm, llm_request
+):
+  """Test that live connections preserve thinking_config from generate config."""
+  thinking_config = types.ThinkingConfig(
+      thinking_budget=10,
+      include_thoughts=True,
+  )
+  llm_request.config.thinking_config = thinking_config
+  llm_request.live_connect_config = types.LiveConnectConfig()
+
+  mock_live_session = mock.AsyncMock()
+
+  with mock.patch.object(gemini_llm, "_live_api_client") as mock_live_client:
+
+    class MockLiveConnect:
+
+      async def __aenter__(self):
+        return mock_live_session
+
+      async def __aexit__(self, *args):
+        pass
+
+    mock_live_client.aio.live.connect.return_value = MockLiveConnect()
+
+    async with gemini_llm.connect(llm_request):
+      mock_live_client.aio.live.connect.assert_called_once()
+      config_arg = mock_live_client.aio.live.connect.call_args.kwargs["config"]
+      assert config_arg.thinking_config == thinking_config
+
+
 @pytest.mark.parametrize(
     (
         "api_backend, "