When are you integrating Kimi K2 thinking? The only thing holding Maple back is how outdated the models are.

Reply to this note

Please Login to reply.

Discussion

Just launched DeepSeek Terminus v3.1.

Always evaluating the latest models. So many come and go that we take our time. In the AI world, a long time means "weeks" 😆

Do you guys run the models on your own hardware ?

We use confidential computing providers and chain together the cryptographic attestations.