Moonshot releases Kimi K2.6 — current state-of-the-art for open-weights coding
Moonshot AI shipped Kimi K2.6, a coding-specialized open-weights model that posts the strongest SWE-bench Verified score among open releases — narrowly ahead of DeepSeek V4 Pro on multi-file edits.
K2.6 uses a mixture-of-experts architecture and is released under a permissive license that allows commercial use. Weights are on Hugging Face. The release continues the Chinese open-source labs' aggressive pace through 2026.
Independent reproductions from third parties are confirming the headline numbers — important since several earlier Chinese-lab releases over-fit to their own benchmarks. Early Cursor and Continue users are reporting it as a viable Claude/GPT replacement for code-only workflows.