Kimi K2.7 Code Is Now in GitHub Copilot and It's the First Open-Weight Model to Make the Cut

GitHub added Kimi K2.7 Code to the Copilot model picker on July 1, and the framing matters: it’s the first open-weight model to reach general availability in the platform.

The model comes from Moonshot AI. Kimi K2.7 Code is an open-weight coding model, which means users and organizations can run it elsewhere and inspect the weights. For GitHub Copilot, that also means it runs at a lower cost than the proprietary options, which GitHub says was part of the reasoning.

Rollout and plan availability

Kimi K2.7 Code is rolling out first to Copilot Pro, Pro+, and Max plans. Business and Enterprise customers get it a few weeks later. For those two tiers, the model is off by default — administrators have to enable it explicitly. GitHub’s guidance is that org admins should evaluate open-weight models against their security and compliance requirements before flipping the switch.

If you’re on a Pro or Pro+ plan, you can select it from the model picker in VS Code 1.127 or later, as well as in Visual Studio, JetBrains, Xcode, and Eclipse.

Billing

It runs under usage-based billing at provider list pricing, hosted on Microsoft Azure. That pricing is separate from your standard Copilot plan credits, the same structure GitHub uses for other models added to the picker via third-party providers.

Why it matters

Most of the models in the Copilot picker are closed: GPT variants, Claude, Gemini. Adding an open-weight option changes the calculus for teams that care about auditability, portability, or running models in their own infrastructure for sensitive code. Whether Kimi K2.7 Code performs well enough to be a daily driver for serious coding tasks is something users will have to judge from their own usage, but the fact that GitHub is opening the door to open-weight models suggests more may follow.

Source: github.blog