The change makes sense, considering our current approach of sticking to the imposed OpenAI "standard".
It's not a blocker but as followups would be nice to:
- test against the real Anthropic API with cache settings
- have a discussion about the option of allowing users to access the raw provider responses
Originally posted by @daavoo in #750 (review)
The change makes sense, considering our current approach of sticking to the imposed OpenAI "standard".
It's not a blocker but as followups would be nice to:
Originally posted by @daavoo in #750 (review)