fix(core): backfill OpenAI responses model in spans

[meronogbai]

Summary

OpenAI Responses API calls can omit model in the request payload when using stored prompts. In that case, Sentry currently records gen_ai.request.model = "unknown" even though the final OpenAI response includes the actual model.

This change keeps the existing request-time fallback behavior, but overwrites the request model from response.model once the Responses API response arrives. It also updates the span name from chat unknown to chat .

This is now handled for both:

• non-streaming Responses API calls
• streaming Responses API calls

Agents dashboard screenshots

In production where this change isn't applied:

In development where this change was applied manually:

---

Before submitting a pull request, please take a look at our
Contributing guidelines and verify:

• If you've added code that should be tested, please add tests.
• Ensure your code lints and the test suite passes (yarn lint) & (yarn test).
• Link an issue if there is one related to your pull request. If no issue is linked, one will be auto-generated and linked.

Closes #issue_link_here

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

[nicohrubec]

Hey, this approach doesn't really make sense to me. If there is no model defined at request time by the user then it seems to me semantically incorrect to backfill the request model retrospectively. Can you maybe explain the use case here a bit more and what exactly you are trying to achieve? If your concern is mainly that the UI shows the actually used model then this is potentially something that should be fixed in the product instead of the SDK.

[meronogbai]

The model is defined in openai's prompt dashboard. So it's defined elsewhere but not in the code calling the openai sdk.

See example in

https://developers.openai.com/api/docs/guides/prompting#create-a-prompt

My concern is the UI, as you said. I'm concrned about different models being under "unknown" . Can you point to where I should fix it if not here?