Purpose of response.reasoning_text

API
1

https://platform.openai.com/docs/api-reference/responses_streaming/response/reasoning_text/delta

Based on my understanding from the official OpenAI documentation, I believe that for reasoning models, raw tokens are not provided. Instead, one should request a “reasoning summary” to obtain the desired value.

However, upon examining the Response object in the Response API, I’ve noticed a response specification that includes reasoning_text.

I’m curious about the purpose of this object and how it can be accessed or called.

If anyone has information about this, please let me know.
Thank you!

2 Likes
2
1 Like
3

The gpt-oss model offers full access, I missed it. Thank you for sharing.

4

How should we be using response.reasoning_text in stream?

Looking at the API, if it were a reasoning.summary_text item, the stream would look like":

ReasoningSummaryPartAdded

ReasoningSummaryTextDelta

ReasoningSummaryTextDelta

…

ReasoningSummaryTextDone

ReasoningSummaryPartDone

But in GPTOSS, we could get ResponseReasoningText. So it should look like

ResponseContentPartAdded

ResponseReasoningTextDelta

ResponseReasoningTextDelta

…

ResponseReasoningTextDone

BUT, ResponseContentPartDone does not take the type ResponseReasoningItem, it’s only

Part: TypeAlias = Annotated[Union[ResponseOutputText, ResponseOutputRefusal], PropertyInfo(discriminator=“type”)]