GitHub has made Kimi K3, an open-weight large language model, generally available inside GitHub Copilot across nearly every surface developers touch it from: VS Code, Visual Studio, JetBrains, Xcode, Eclipse, the Copilot CLI, the Copilot cloud agent, github.com, and the mobile apps. The rollout briefly paused because of an unrelated GitHub Actions incident, then resumed on August 6. What makes this notable is not just another model option in a picker, it is the growing normalization of open-weight models sitting alongside proprietary ones from OpenAI, Anthropic, and Google inside the dominant commercial coding assistant. Kimi K3 is hosted by GitHub via Fireworks AI rather than run by the model's original developer, and it is priced competitively under Copilot's usage-based billing at $3 per 1M input tokens, $15 per 1M output tokens, and $0.30 per 1M cached input tokens, notably cheaper than many frontier proprietary alternatives while still claiming frontier-level agentic coding ability. For individual Pro, Pro+, and Max subscribers the model shows up automatically as rollout reaches them; for Business and Enterprise customers it ships off by default, requiring an admin to explicitly enable an open-weight-model policy after reviewing it against their own security, compliance, and data-governance posture. That opt-in gate signals GitHub treating open-weight models as a distinct risk and governance category from its default vendor-hosted models, likely to become a template as more open-weight coding models enter the assistant market through 2026. For engineering teams, the practical takeaways are cost arbitrage for high-volume agentic coding tasks and the need to update internal AI-tool-approval processes to account for a growing menu of open-weight options inside tools they already trust.