Atlas

Benchmarks

← All benchmarks

Kimi Code Bench v2

Code · 2026-06-11

In-house coding-agent benchmark for production software tasks.

Top models (higher is better)

ModelScore
Claude Fable 576.9
Kimi K372.9
Opus 4.871.7
GPT-5.569.0
GPT-5.6 Sol64.8
GLM-5.264.2
Loading Atlas data…