Is there any way to get back usage tokens from responses?

[bradenkinard]

Really liking the library, its very slick. However, I'd like to be able to get back the usage tokens OpenAI returns in the response, or more generally, the metadata beyond just the response message. The Pydantic validation and automatic retries this library offers is incredibly useful, but once that is done, having access to the raw OpenAI response would also be nice, at least in my case.

[jxnl]

its not very documented but Model.resp._raw_response is the raw response. I'll improve the documentation there:

[bradenkinard]

Perfect! This is exactly what I needed. Thanks!