OpenAI has published new prompt templates for GPT-6.2 Astra and Sol after observing overly brief answers, unnecessary follow-up questions and excessive testing in coding tasks. Developers migrating from older models are advised to review instructions such as “be concise,” because Astra already writes more directly by default.

The guidance tells models to make reversible assumptions and continue working when authorized, pausing only for a genuine blocker, major scope change or destructive action. For code, OpenAI recommends reusing existing project logic, avoiding needless abstractions and adding tests only when requested, required by the repository or needed for a specific regression.

The Responses API also changes how conversation edits work. Altering history, switching to certain models, or changing the user or organization can remove earlier reasoning content, potentially increasing token use when work must be repeated. The usual `temperature` and `top_p` controls are no longer supported. OpenAI recommends medium reasoning effort for new integrations and low effort for existing applications that previously used none or minimal. Teams should test both behavior and cost before replacing an older model.