ResponsesAPIStreamingResponse

litellm/types/llms/openai.py

@@ -756,3 +756,8 @@ class ResponsesAPIResponse(TypedDict, total=False):
     truncation: Optional[Literal["auto", "disabled"]]
     usage: Optional[ResponseUsage]
     user: Optional[str]
+
+
+class ResponsesAPIStreamingResponse(TypedDict, total=False):
+    type: str
+    response: ResponsesAPIResponse

tests/llm_responses_api_testing/test_openai_responses_api.py

@@ -16,3 +16,16 @@ async def test_basic_openai_responses_api():
     print("litellm response=", json.dumps(response, indent=4, default=str))
 
     # validate_responses_api_response()
+
+
+@pytest.mark.asyncio
+async def test_basic_openai_responses_api_streaming():
+    litellm._turn_on_debug()
+    response = await litellm.aresponses(
+        model="gpt-4o",
+        input="Tell me a three sentence bedtime story about a unicorn.",
+        stream=True,
+    )
+
+    for event in response:
+        print("litellm response=", json.dumps(event, indent=4, default=str))