We’re a tier 5 OpenAI user since April 2023. We were the one of biggest publishers of ChatGPT plugins and then build multiple AI apps (1, 2, 3 and others) powered exclusively by the OpenAI API. Everything has been great so far, but with the recent development of reasoning AI models, we strongly feel the need to access reasoning content through the API.
There are two main use cases where we need this:
- When debugging prompts, the reasoning content helps us understand how the LLM “thinks” and makes decisions. For example, if the flow of thoughts is incorrect, we know something needs to be adjusted, either the prompt or the context provided.
- We also want to display this information to users in our apps when certain actions are taken, as we believe it offers transparency and a helpful way to communicate what’s happening behind the scenes.
So I’m wondering—are there any plans to add this field to the API response?