Moonshot AI open-sources Kimi K2.7 Code, an agentic coding model

China's Moonshot AI has released Kimi K2.7 Code, an open-weight agentic coding model, publishing the full weights on Hugging Face under a permissive Modified MIT license. Built on the trillion-parameter K2.6 mixture-of-experts design (32B active, 256K context), it targets long-horizon software tasks. Moonshot says it beats K2.6 on the firm's own coding benchmarks — +21.8% on one — while using about 30% fewer reasoning tokens.
3 sources
- 01MarkTechPost — Moonshot AI Releases Kimi K2.7-Code: a Coding Model Reporting +21.8% on Kimi Code Bench v2 Over K2.6marktechpost.com
- 02DevOps.com — Moonshot AI's Kimi K2.7-Code Targets Token Efficiency in Agentic Codingdevops.com
- 03Cloudflare — Moonshot AI Kimi K2.7 Code now available on Workers AIdevelopers.cloudflare.com