GitHub Copilot can use AI models from multiple providers, but there is no single model roster that applies to every user. The models you can select depend on your Copilot plan, the client you use, and any organization or enterprise policies. GitHub documents supported models and selection options; its public pages do not explain every detail of the infrastructure or request routing behind each Copilot feature.
Which AI models does GitHub Copilot support?
GitHub’s supported-models reference lists models from OpenAI, Anthropic, Google, Microsoft, Moonshot AI, and xAI. The roster includes model families such as GPT, Claude, Gemini, MAI-Code, Kimi, and Grok. GitHub says availability can change, so consult the live reference rather than treating any list as permanent.
As GitHub puts it, “GitHub Copilot supports multiple AI models, each with different strengths.” A model appearing in the reference does not mean it is available to every Copilot user or in every Copilot client.
Why does the model picker differ between users?
Model access depends on three things: your subscription, the Copilot client you are using, and model policies configured by your organization or enterprise. GitHub’s documentation on supported models and model access controls explains these availability constraints. In particular, Copilot Free and Copilot Student users can use models through Auto model selection only; they do not have the same manual model choices as eligible paid-plan users.
Recommended Free Tools
If a model is missing from your picker, check that it is supported in your client and plan, then ask your organization administrator whether a policy restricts it. A model’s presence on GitHub’s overall support page alone does not establish that your account can select it.
What does Auto model selection do?
GitHub says Auto selects a model based on availability and task complexity. The selection therefore operates within the models accessible in your circumstances; it does not override plan, client, or organization restrictions. GitHub also states that paid Copilot plan users qualify for a discount on model costs when using Auto. Check the current model-use guidance for applicable eligibility and cost terms, which may change.
How should you choose a model?
GitHub’s model comparison guidance frames selection around the task and tradeoffs such as speed, cost, accuracy, reasoning, and task-specific strengths. It is guidance for comparing options, not a guarantee that one model will perform best for every prompt or an independent benchmark.
- Quick explanation or small code completion: prioritize responsiveness and whether the model’s capabilities match the narrow task.
- Debugging: consider how much reasoning the problem needs, alongside speed and any cost or AI-credit implications.
- Longer, agentic work: weigh reasoning and task-specific capabilities against time and cost.
- Any task: first confirm the model is available in your plan and client and is not restricted by organization policy.
These are practical criteria for applying GitHub’s comparison framework, not reported performance results. For current model-specific descriptions, use GitHub’s live comparison page rather than assuming family names alone predict how a model will handle your task.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Scan for outdated or missing drivers - takes under a minute3Clear out junk files and repair common Windows errorsWhat is public about Copilot’s model infrastructure?
GitHub publicly documents supported model names and providers, access controls, Auto selection, and comparison guidance. Those pages do not provide a complete account of where each model is hosted or exactly how requests are routed in every Copilot feature. A provider or model listed as supported is not, by itself, evidence of a particular hosting arrangement or internal request path.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
