Skip to main content

Hey, with Kimi K2 release I heard a lot of good responses about using Groq + Kimi for coding agents.

What coding agent you found the most satisfactory and easy to set up to run with Groq?

I’ve seen people use it with Cline and they had good feedbacks. What other options? What about temrinal agents like Claude code?

Please share your experience

While ​@benank loves using Kimi + Cline, I’ve never been a huge fan of Cline and Roo.

I’ve also had mixed results with Claude Code! Everyone around me keeps saying how good Claude Code is but I can never seem to get it to give me good results (with Sonnet 4, maybe that’s the problem). Some are testing Claude Code with Groq/Kimi with a proxy, but again, I’ve for some reason never bought into Claude Code.

I’ve been trying STT.dev’s https://opencode.ai and it seems to be fairly good, but mirrors some of the other prompt/task adherence problems I had with Claude Code.

Maybe it’s just a skill issue…

(though the entire point of agents is to take the human skill out of coding in the first place)

 


> I’ve also had mixed results with Claude Code! 

Actually works great for me (except of reaching daily token limits) but for small-medium CLI-based projects. I would not trust it to pick something about frontend or backend architecture, for this I prefer precise control so Cursor-like IDE is still required.

But if it is NodeJS CLI, Claude Code can easily add new features and test them in realtime, because it’s just simple IO so it can plan the implementation


Oh nice, yeah most things I work on are frontend with some backend API stuff, so it just gets things wrong (while chewing up tons tokens) almost all the time for me.


Reply