Comment by zeroxfe
10 hours ago
That tends to work quite poorly because Claude Code does not use standard completions APIs. I tried it with Kimi, using litellm[proxy], and it failed in too many places.
10 hours ago
That tends to work quite poorly because Claude Code does not use standard completions APIs. I tried it with Kimi, using litellm[proxy], and it failed in too many places.
It worked very well for me using qwen3 coder behind a litellm. Most other models just fail in weird ways though.
opencode is a good alternative that doesnt flake out in this way.