About models in Copilot
GitHub Copilot supports a range of AI models. Different models have different strengths. For example, some prioritize low latency, while others are optimized for complex reasoning or large context windows. For a full list of supported models, see GitHub Copilot中支持的 AI 模型.
Depending on the surface you're using, your plan, and any policies your organization or enterprise has set, you can often choose which model handles your request, or let auto model selection choose one for you. Model choice affects the speed, cost, and quality of your results, so understanding your options helps you get the most out of every feature.
In addition, Copilot also has utility models that run automatically in the background, and for Copilot Business and Copilot Enterprise, base and long-term support (LTS) models for extended stability.
Model choice
When you're using Copilot, whether that's on GitHub, Copilot CLI, the GitHub Copilot app and supported IDEs, you can select the model that best suits your task. For guidance on which model fits common tasks, such as general-purpose coding, large-scale refactors, or documentation writing, see AI 模型比较.
What determines which models are available to you
The models you see in Copilot depend on several factors:
- Your plan. Some models are available on all plans, while others, such as premium models, require a paid plan. See GitHub Copilot计划.
- The Copilot surface or client you're using. Not every model is available in every IDE, on GitHub, in Copilot CLI, or in the GitHub Copilot app. See GitHub Copilot中支持的 AI 模型.
- Administrator policies. If you're part of an organization or enterprise, administrators can enable or disable specific models, which affects what you and auto model selection can use. See 在 GitHub Copilot 中配置对 AI 模型的访问.
- Model availability over time. Providers periodically release new models and retire older ones, so the specific models available to you can change. Check GitHub Copilot中支持的 AI 模型 for the current list.
Auto model selection
Rather than requiring you to track every model's strengths yourself, you can select Auto in the model picker. Auto model selection then routes your request to an appropriate supported model based on the complexity of your task and real-time model availability. See 关于 Copilotauto model selection.
Bring your own key
With bring your own key (BYOK), you (or your enterprise) can connect a custom or provider-hosted model to Copilot. Availability depends on the mechanism you use and the client you're working in. See GitHub Copilot 使用您自己的密钥.
Utility models for background features
Some Copilot features, such as generating a commit message or a chat session title, run on small utility models that work automatically in the background. Utility models aren't part of the model picker, and you can't select them directly. See 实用工具模型.
Base and long-term support models
Copilot Business and Copilot Enterprise accounts have access to two additional model designations:
- A base model is used automatically when no other model is enabled for the account.
- A long-term support (LTS) model is supported for a defined extended period of time, currently one year from its designation date.
See 基础和长期支持 （LTS） 模型.