LiveSWEBench: New benchmark from creators of LiveBench

[bogdan0083]

Basically it's a benchmark for code agents: https://liveswebench.ai/

Might be interesting to test Kilo Code on that one.

[ofou]

That’s a great idea, @bogdan0083! I believe it's very important to do evaluations frequently. We’ll begin working on this once we feel more comfortable with our product and have fixed some stuff we're working on right now.

What other evals do you think might be relevant for us? I think about Aider Polyglot too, I talked a bit about them in our blog, let me know if you have any other eval ideas!

[bogdan0083]

Well, I'm not really competent in this area to suggest anything, but I believe there's also SWEBench Verified, it's also worth checking out for sure!