weiqingy opened a new issue, #905: URL: https://github.com/apache/flink-agents/issues/905
### Search before asking - [X] I searched in the [issues](https://github.com/apache/flink-agents/issues) and found nothing similar. ### Description The OpenAI Chat Completions chat model defaults to `gpt-3.5-turbo` when the caller omits `model`. This is a legacy OpenAI model — it has been superseded by cheaper, more capable small models, and OpenAI itself steers new integrations away from it. A user who takes the default today silently gets an outdated model. The stale default lives in both language runtimes and the docs, so a user hits it regardless of which API they use: - Java — `OpenAICompletionsSetup.DEFAULT_MODEL` (`integrations/chat-models/openai/src/main/java/org/apache/flink/agents/integrations/chatmodels/openai/OpenAICompletionsSetup.java:61`) - Python — `DEFAULT_OPENAI_MODEL` (`python/flink_agents/integrations/chat_models/openai/openai_chat_model.py:38`) - Docs — the `OpenAICompletionsSetup` parameter tables for both languages (`docs/content/docs/development/chat_models.md`) The Java Javadoc example already uses `gpt-4o-mini` (`OpenAICompletionsSetup.java:45`), so the code base is already inconsistent with its own default. Proposed fix: update the default in both runtimes and the docs to `gpt-4o-mini` — the current low-cost successor to `gpt-3.5-turbo`, and the value the Javadoc example already demonstrates. This keeps a sensible zero-config default while dropping the legacy model, and it keeps Java and Python in parity. ### Are you willing to submit a PR? - [X] I'm willing to submit a PR! -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
