AI model · by Moonshot AI
Kimi K2.7 Code
Kimi K2.7 Code is Moonshot AI's agent-style model aimed at coding, built on Kimi K2.6 for lengthy software engineering work.
- 262K tokens context
- Open weights · Modified Mit licence
- Checked against Moonshot AI's page 19 Oct 2026
What Kimi K2.7 Code is for
Moonshot AI says its Kimi K2.7 Code model, derived from Kimi K2.6, is geared toward coding and autonomous agent tasks, with notable gains on practical extended programming jobs and on taking multi-stage software projects through to completion. Developers can reach it through Moonshot's API at platform.moonshot.ai, which supports OpenAI-style and Anthropic-style interfaces, and the company suggests serving it with vLLM, SGLang, or KTransformers.
What Moonshot AI says it does well
- Improved performance on real-world long-horizon coding tasks, strengthening end-to-end task completion across complex software engineering workflows
- Improved token efficiency, reducing thinking-token usage by approximately 30% compared with Kimi K2.6
- Supports image and video input
- Preserve thinking mode retains full reasoning content across multi-turn interactions, enhancing performance in coding agent scenarios
Where to use it
- Maker's official API on platform.moonshot.ai, offered with OpenAI/Anthropic-compatible endpoints
- Inference engines vLLM, SGLang, and KTransformers