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

sentry_sdk/integrations/openai.py

@@ -612,6 +612,8 @@
     def new_iterator() -> "Iterator[ChatCompletionChunk]":
         nonlocal ttft
         for x in old_iterator:
+            span.set_data(SPANDATA.GEN_AI_RESPONSE_MODEL, x.model)
+
             with capture_internal_exceptions():
                 if hasattr(x, "choices"):
                     choice_index = 0
@@ -654,6 +656,8 @@
     async def new_iterator_async() -> "AsyncIterator[ChatCompletionChunk]":
         nonlocal ttft
         async for x in old_iterator:
+            span.set_data(SPANDATA.GEN_AI_RESPONSE_MODEL, x.model)
+
             with capture_internal_exceptions():
                 if hasattr(x, "choices"):
                     choice_index = 0

tests/integrations/openai/test_openai.py

@@ -511,6 +511,8 @@ def test_streaming_chat_completion_no_prompts(
     span = tx["spans"][0]
     assert span["op"] == "gen_ai.chat"
 
+    assert span["data"][SPANDATA.GEN_AI_RESPONSE_MODEL] == "model-id"
+
     assert SPANDATA.GEN_AI_SYSTEM_INSTRUCTIONS not in span["data"]
     assert SPANDATA.GEN_AI_REQUEST_MESSAGES not in span["data"]
     assert SPANDATA.GEN_AI_RESPONSE_TEXT not in span["data"]
@@ -656,6 +658,8 @@ def test_streaming_chat_completion(sentry_init, capture_events, messages, reques
             },
         ]
 
+    assert span["data"][SPANDATA.GEN_AI_RESPONSE_MODEL] == "model-id"
+
     assert "hello" in span["data"][SPANDATA.GEN_AI_REQUEST_MESSAGES]
     assert "hello world" in span["data"][SPANDATA.GEN_AI_RESPONSE_TEXT]
 
@@ -762,6 +766,8 @@ async def test_streaming_chat_completion_async_no_prompts(
     span = tx["spans"][0]
     assert span["op"] == "gen_ai.chat"
 
+    assert span["data"][SPANDATA.GEN_AI_RESPONSE_MODEL] == "model-id"
+
     assert SPANDATA.GEN_AI_SYSTEM_INSTRUCTIONS not in span["data"]
     assert SPANDATA.GEN_AI_REQUEST_MESSAGES not in span["data"]
     assert SPANDATA.GEN_AI_RESPONSE_TEXT not in span["data"]
@@ -897,6 +903,8 @@ async def test_streaming_chat_completion_async(
     span = tx["spans"][0]
     assert span["op"] == "gen_ai.chat"
 
+    assert span["data"][SPANDATA.GEN_AI_RESPONSE_MODEL] == "model-id"
+
     param_id = request.node.callspec.id
     if "blocks" in param_id:
         assert json.loads(span["data"][SPANDATA.GEN_AI_SYSTEM_INSTRUCTIONS]) == [