completely anecdote: vision seems > gemini 2.5 but less than 3.0

I haven't used it much for programming, but it feels like a model 6 months out of date for general use

I've been using kimi, though not kimiclaw, for research and it is good - comparable to phind, better than GLM 4.7 . Opus 4.6 wasn't as good for my particular domain of interest. I think the long term pricing asymptote for US vs china is essentially dependent on energy pricing and so china will continue to undercut US AI pricing.

Did you scroll through the pricing options? The largest Kimi plan is $199/month. “Much better” depends on how much usage is included vs. Anthropic plans/API costs.

Opus is definitely in its own league. I use Kimi/Gemini-cli code regularly to save cost and from my experience, Kimi 2.5 is more solid than Gemini Flash 3.0 for coding. While Gemini Flash 3.0 is generally faster, it usually break the syntax and skip important prompt. Kimi 2.5 can write very good code and can plan very well.

Kimi 2.5 is a great model. Especially for writing, understanding many different text sources, reasoning etc. I haven't personally used it for coding.