Open weights walk in the front door of GitHub Copilot
Kimi K2.7 Code is the first open weight model in Copilot's picker. The story is distribution, not benchmarks.
GitHub Copilot now offers Moonshot’s Kimi K2.7 Code as a selectable model: the first open weight model in the picker, alongside models from OpenAI, Anthropic, Google, and Microsoft. It is hosted by GitHub on Azure, billed at provider list pricing, and rolling out from individual plans toward Business and Enterprise.
The significance is distribution. Open weight models have been credible on coding work for a while; what they lacked was a procurement path that did not require anyone to build serving infrastructure or onboard a new vendor. Appearing inside a tool enterprises already license removes that barrier entirely. K2.7 Code is a 1 trillion parameter mixture of experts with 32 billion active per token, which is how it comes in cheaper than closed peers on list price.
The adoption note: it ships off by default for Copilot Business and Enterprise. An administrator has to enable the policy before anyone can select it. Treat that switch as a decision, not a formality. Run the model against your own eval set first, then enable it with usage reporting on, and let displacement data tell you what your developers actually prefer.
Written by Adib Kadir. Product and engineering executive focused on rolling out AI at enterprise scale.
Start a conversation