Kimi K2.7 Code becomes first open-weight model in GitHub Copilot's picker

Kimi K2.7 Code, an open-weight coding model from China’s Moonshot AI, became generally available in GitHub Copilot on July 1, 2026, GitHub said in its changelog. GitHub calls it the first open-weight model selectable in the Copilot model picker, hosted by GitHub on Microsoft Azure.

The move brings a Chinese open-weight model into Microsoft’s flagship developer tool, giving Copilot’s paying users a self-hostable alternative to the proprietary models that have dominated the assistant.

Kimi K2.7 Code rolls out immediately to Copilot Pro, Pro+ and Max plans, billed at provider list pricing under usage-based billing. Business and Enterprise access comes later, is off by default, and requires an administrator policy to enable. The model is available across VS Code, Visual Studio, the Copilot CLI, JetBrains IDEs, Xcode, Eclipse and GitHub.com.

The model uses a mixture-of-experts design with 1 trillion total parameters and 32 billion activated across 384 experts, a 256K-token context window, and is released under a modified MIT license, per Moonshot AI’s model card. Moonshot AI says the model scores 62.0 on its own Kimi Code Bench v2, up from 50.9 for the prior K2.6 generation, and reduces thinking-token usage by about 30% versus K2.6.

Those figures come from Moonshot AI’s in-house benchmark and have not been independently verified; a vendor’s own eval is a claim, not a settled result. GitHub hosting the weights on Azure also sidesteps the data-residency and trust concerns that have kept some enterprises wary of Chinese models running on third-party infrastructure.

Whether other open-weight models follow Kimi into the Copilot picker, or it remains a one-off, will show how far GitHub intends to open its assistant beyond proprietary systems.