GitHub has made Copilot Fusion generally available. Developers can select it in Copilot’s model picker to have the system choose and coordinate AI models for each prompt. They choose an effort level instead of picking a model for each task.
Fusion combines the multi-model approach from Project HydraFusion with Copilot’s automatic model selection, according to GitHub. Standard is intended for everyday coding, High for challenging tasks and deeper reasoning, and Max for the most demanding tasks.
GitHub says Fusion can use one model to answer directly. It can also have one model draft an answer and check whether a stronger model is needed, or have a second model review a draft before the first revises it.
Read nextMicrosoft makes its Copilot agent harness available through FoundryAdministrators can control model access
Fusion uses only models enabled under an enterprise’s or organization’s model policies. Enterprise administrators can set each model’s policy for Fusion separately and exclude a model from Fusion without removing it for other uses. They can also make any Fusion effort level the default through Enterprise managed settings.
GitHub says Fusion Max scored 60 in the Artificial Analysis Coding Agent Index, within two points of Claude Code with Opus 5.5 Max. The company reports that Fusion Max achieved its result at about half the average API cost per task, with about half as many tokens and about 40% less time per task.
Paid users get a temporary discount
Users are billed for every model Fusion uses at that model’s standard rate. For a limited time, GitHub says users on paid Copilot plans receive a 30% discount on model costs when using Fusion.
Copilot Fusion is available on all Copilot plans in Visual Studio Code, the GitHub Copilot app, GitHub Copilot CLI and JetBrains IDEs. GitHub says support for additional clients is coming soon. Users can update their client, select Fusion in the model picker and choose an effort level.
