fix(openai): Attach response model with streamed Responses API

Author: alexander-alderman-webb

sentry_sdk/integrations/openai.py

@@ -760,6 +760,8 @@ def new_iterator() -> "Iterator[ChatCompletionChunk]":
                     data_buf[0].append(x.delta or "")
 
                 if isinstance(x, ResponseCompletedEvent):
+                    span.set_data(SPANDATA.GEN_AI_RESPONSE_MODEL, x.response.model)
+
                     _calculate_token_usage(
                         input,
                         x.response,
@@ -807,6 +809,8 @@ async def new_iterator_async() -> "AsyncIterator[ChatCompletionChunk]":
                     data_buf[0].append(x.delta or "")
 
                 if isinstance(x, ResponseCompletedEvent):
+                    span.set_data(SPANDATA.GEN_AI_RESPONSE_MODEL, x.response.model)
+
                     _calculate_token_usage(
                         input,
                         x.response,

tests/integrations/openai/test_openai.py

@@ -2460,6 +2460,7 @@ async def test_ai_client_span_streaming_responses_async_api(
 
     expected_data = {
         "gen_ai.operation.name": "responses",
+        "gen_ai.response.model": "response-model-id",
         "gen_ai.response.streaming": True,
         "gen_ai.system": "openai",
         "gen_ai.response.time_to_first_token": mock.ANY,