diff --git a/sentry_sdk/integrations/google_genai/streaming.py b/sentry_sdk/integrations/google_genai/streaming.py
index 86cdcf29ba..64d5a354d2 100644
--- a/sentry_sdk/integrations/google_genai/streaming.py
+++ b/sentry_sdk/integrations/google_genai/streaming.py
@@ -59,6 +59,11 @@ def accumulate_streaming_response(
     model = None
 
     for chunk in chunks:
+        # Keep the last non-empty value, so a chunk without these fields
+        # doesn't clear one seen earlier.
+        response_id = getattr(chunk, "response_id", None) or response_id
+        model = getattr(chunk, "model_version", None) or model
+
         # Extract text and tool calls
         if getattr(chunk, "candidates", None):
             for candidate in getattr(chunk, "candidates", []):
diff --git a/tests/integrations/google_genai/test_google_genai.py b/tests/integrations/google_genai/test_google_genai.py
index 444333e45f..997162b329 100644
--- a/tests/integrations/google_genai/test_google_genai.py
+++ b/tests/integrations/google_genai/test_google_genai.py
@@ -866,6 +866,14 @@ def test_streaming_generate_content(
         assert (
             chat_span["attributes"][SPANDATA.GEN_AI_REQUEST_MODEL] == "gemini-1.5-flash"
         )
+        assert (
+            chat_span["attributes"][SPANDATA.GEN_AI_RESPONSE_ID]
+            == "response-id-stream-123"
+        )
+        assert (
+            chat_span["attributes"][SPANDATA.GEN_AI_RESPONSE_MODEL]
+            == "gemini-1.5-flash"
+        )
     else:
         events = capture_events()
 
@@ -916,6 +924,10 @@ def test_streaming_generate_content(
 
         # Verify model name
         assert chat_span["data"][SPANDATA.GEN_AI_REQUEST_MODEL] == "gemini-1.5-flash"
+        assert (
+            chat_span["data"][SPANDATA.GEN_AI_RESPONSE_ID] == "response-id-stream-123"
+        )
+        assert chat_span["data"][SPANDATA.GEN_AI_RESPONSE_MODEL] == "gemini-1.5-flash"
 
 
 @pytest.mark.parametrize("span_streaming", [True, False])