Moonshot AI releases Kimi K2.7 Code open-weights

Jun 12, 2026

Moonshot AI has released Kimi K2.7 Code, an open-weights model priced at a fraction of competitors like GPT-5.5 and Claude Fable 5. At $4.00 per million output tokens, it is up to 12x cheaper than Anthropic's top model. While Moonshot claims performance gains, the release lacks independent benchmarks, and researchers find it "more honest but not more capable" than its predecessor.

Kimi K2.7 Code release

▪Kimi K2.7 Code features a 256,000-token context window and is multimodal, processing text, images, and video.
▪The model's license requires commercial products with over 100M monthly users or $20M in monthly revenue to display "Kimi K2.7 Code".
▪The model's release lacked independent, third-party benchmark scores, with Moonshot AI publishing only its own results.
▪Moonshot AI claims Kimi K2.7 Code uses approximately 30% fewer "thinking tokens" than its predecessor, K2.6.
▪The model has a 1 trillion-parameter Mixture-of-Experts (MoE) architecture, with 32 billion parameters active per token.
▪In Moonshot's own benchmarks, Kimi K2.7 Code trails GPT-5.5 on most coding tests.
▪On the MCPMark Verified agent benchmark, Kimi K2.7 Code scored 81.1, beating Claude Opus 4.8's score of 76.4.
▪Kimi K2.7 Code is available on Hugging Face under a Modified MIT license.
▪A researcher's test on the KernelBench-Hard benchmark found Kimi K2.7 Code to be "more honest but not more capable" than K2.6.
▪A "HighSpeed inference mode" for Kimi K2.7 Code was announced, claiming to be approximately 6 times faster.
▪Moonshot AI, a Beijing-based lab, released the open-weights model Kimi K2.7 Code on June 12, 2026.

Proprietary model pricing

▪Anthropic's Claude Opus 4.8 costs $5.00 per million input tokens and $25.00 per million output tokens.
▪OpenAI's GPT-5.5 is priced at $5.00 per million input tokens and $30.00 per million output tokens.
▪Anthropic's Claude Fable 5 is priced at $10.00 per million input tokens and $50.00 per million output tokens.
▪On June 12, 2026, the U.S. government issued an export control directive forcing Anthropic to suspend access to Claude Fable 5 worldwide.

Open-source LLM cost comparison

▪GPT-5.5's output pricing is 7.5 times more expensive than Kimi K2.7 Code's.
▪The price for cached input tokens for Kimi K2.7 Code is $0.19 per million.
▪Based on output pricing, Claude Fable 5 is more than 12 times as expensive as Kimi K2.7 Code.
▪Kimi K2.7 Code's API is priced at $0.95 per million input tokens and $4.00 per million output tokens.