Moonshot AI Releases Kimi K2.7-Code Open Weights on Hugging Face

MarkTechPost·high signal

Moonshot AI dropped Kimi K2.7-Code on Hugging Face (June 12) — a 1T-parameter MoE (32B active, 384 experts) with a 256K context window under a Modified MIT license, specialized for long-horizon agentic software engineering. Moonshot reports +21.8% on Kimi Code Bench v2, +11.0% on Program Bench, +31.5% on MLS Bench Lite, and ~10% agentic gains on MCP Atlas/MCP Mark Verified versus K2.6 — while using roughly 30% fewer 'thinking' tokens. The token-efficiency claim is the builder-relevant hook: cheaper agentic runs if it holds up.