test(e2e): gpt-5.5 drops unsupported temperature with drop_params

tests/e2e/coverage_registry/llm_conversational.yaml

@@ -9,6 +9,7 @@
 - {id: llm.chat_completions.openai.service_tier.nonstream.works, module: llm, tier: P1, subject_endpoint: chat_completions, route: openai, capability: service_tier, streaming: nonstream, assertions: [works], source: "OpenAI service_tier param", rationale: "OpenAI scale-tier request option is forwarded and echoed"}
 - {id: llm.chat_completions.openai.thinking.nonstream.works, module: llm, tier: P1, subject_endpoint: chat_completions, route: openai, capability: thinking, streaming: nonstream, assertions: [works], source: "model_prices json", rationale: "o-series reasoning; emerging"}
 - {id: llm.chat_completions.openai.structured_output.nonstream.works, module: llm, tier: P1, subject_endpoint: chat_completions, route: openai, capability: structured_output, streaming: nonstream, assertions: [works], source: "model_prices json", rationale: "response_schema extraction"}
+- {id: llm.chat_completions.openai.drop_params.nonstream.works, module: llm, tier: P1, subject_endpoint: chat_completions, route: openai, capability: drop_params, streaming: nonstream, assertions: [works], source: "llms/openai/chat/gpt_5_transformation.py", rationale: "drop_params must drop temperature != 1 for gpt-5 models whose default reasoning_effort is active (gpt-5.5 defaults to medium), else the provider 400s (LIT-3797)", fail_before_fix: proven}
 - {id: llm.chat_completions.anthropic.basic.nonstream.works, module: llm, tier: P0, subject_endpoint: chat_completions, route: anthropic, capability: basic, streaming: nonstream, assertions: [works], source: "proxy_server.py:8455", rationale: "P0 route translated to Anthropic"}
 - {id: llm.chat_completions.anthropic.basic.stream.works, module: llm, tier: P0, subject_endpoint: chat_completions, route: anthropic, capability: basic, streaming: stream, assertions: [works], source: "proxy_server.py:8455", rationale: "Streaming translation"}
 - {id: llm.chat_completions.anthropic.tool_use.nonstream.works, module: llm, tier: P0, subject_endpoint: chat_completions, route: anthropic, capability: tool_use, streaming: nonstream, assertions: [works], source: "model_prices json", rationale: "Claude tool_use; high usage"}

tests/e2e/coverage_registry/schema.py

@@ -58,6 +58,7 @@ LlmCapability = Literal[
     "assume_role",
     "basic",
     "count_tokens",
+    "drop_params",
     "long_context_1m",
     "mid_conversation_system",
     "pdf_input",

tests/e2e/llm_translation/test_chat_completions_regression_e2e.py

@@ -81,6 +81,54 @@ class TestChatCompletionsRegression:
         ), f"{model} ({route}): 200 with an empty completion (#28991): {response}"
 
 
+class TestGpt5TemperatureDropParams:
+    """gpt-5 family models whose default reasoning_effort is active (gpt-5.5
+    defaults to medium, unlike gpt-5.1 which defaults to none) reject
+    temperature != 1. A deployment with drop_params must drop the temperature
+    instead of forwarding it and surfacing the provider's 400 (LIT-3797)."""
+
+    @pytest.mark.covers(
+        "llm.chat_completions.openai.drop_params.nonstream.works",
+        exercised_on=["chat_completions"],
+    )
+    def test_gpt55_drops_unsupported_temperature(
+        self, client: PassthroughClient, resources: ResourceManager
+    ) -> None:
+        model = f"e2e-gpt55-temp-drop-{unique_marker()}"
+        model_id = client.proxy.create_model(
+            model,
+            LiteLLMParamsBody(
+                model="openai/gpt-5.5",
+                api_key="os.environ/OPENAI_API_KEY",
+                drop_params=True,
+            ),
+        )
+        resources.defer(lambda: client.proxy.delete_model(model_id))
+        key = resources.key()
+
+        response = unwrap(
+            client.proxy.chat(
+                key,
+                ChatBody(
+                    model=model,
+                    messages=[
+                        ChatMessage(
+                            role="user",
+                            content=f"Reply with the single word pong. {unique_marker()}",
+                        )
+                    ],
+                    temperature=0.1,
+                    max_tokens=512,
+                ),
+            )
+        )
+        assert response.choices, f"gpt-5.5 chat returned no choices: {response}"
+        message = response.choices[0].message
+        assert (
+            message is not None and message.content and message.content.strip()
+        ), f"gpt-5.5 with temperature 0.1 returned an empty completion: {response}"
+
+
 class TestCohereChat:
     """Cohere via the OpenAI-compatible /chat/completions path."""
 

tests/e2e/models.py

@@ -172,6 +172,7 @@ class ChatBody(BaseModel):
     messages: list[ChatMessage]
     stream: bool = False
     max_tokens: int | None = None
+    temperature: float | None = None
     user: str | None = None
     metadata: ChatMetadata | None = None
     reasoning_effort: str | None = None
@@ -568,6 +569,7 @@ class LiteLLMParamsBody(BaseModel):
     complexity_router_config: dict[str, object] | None = None
     mock_response: str | None = None
     timeout: float | None = None
+    drop_params: bool | None = None
 
 
 ModelMode = Literal["batch", "realtime", "image_generation"]