GitHub Copilot Fusion Is Live: Let Copilot Pick the Models, Near Claude Code Quality at Half the Cost
GitHub Copilot Fusion is generally available.
It picks and combines models from Anthropic, OpenAI and others for each task, in Standard, High and Max tiers. In an Artificial Analysis evaluation cited by GitHub, Max trails Claude Code with Opus 5.5 by two points at about half the cost per task.
It works on every plan, with 30% off through November 30.
Which model should you use for a given coding task? An expensive one wastes money on easy work, a cheap one stumbles on hard work, and you often forget to switch back. GitHub's answer is to stop choosing and let Copilot do it.
GitHub Copilot Fusion is now generally available. Pick Copilot Fusion in the model picker and it decides, for each task, which models to use and how they work together. You only choose an effort level.
Three tiers, three ways to collaborate
The tiers are Standard, High and Max. Standard favors versatile models for everyday coding, High targets harder tasks that need deeper reasoning, and Max uses the most capable models for the most demanding work.
Behind the scenes, Fusion works in one of three modes. In the simplest case, one model solves the task directly. In Cascade, an efficient model drafts an answer and a quality check decides whether to escalate to a stronger one. In Critique, one model drafts, another reviews, and the original model revises. The approach comes from the earlier research preview HydraFusion, now merged with Copilot's automatic model selection.
Cost is the selling point
GitHub's figures: on the Artificial Analysis Coding Agent Index, Fusion Max scored 60, within two points of Claude Code with Opus 5.5 Max. It did so at about half the average API cost per task, using roughly half the tokens and taking about 40% less time.
Two points for half the price is a good trade for many teams, but this is an evaluation GitHub chose to cite, so real-world results on your own code still need checking. Billing is worth understanding too: you pay each model Fusion uses at its standard rate, so the bill will move with task difficulty. As a limited-time offer, paid-plan users get 30% off model costs when using Fusion through November 30.
What admins can control
Enterprise administrators decide which models Fusion may choose from, can disable a single model for Fusion without affecting anyone else, and can make Fusion the default for everyone.
Fusion is available on all Copilot plans in VS Code, the GitHub Copilot app, Copilot CLI and JetBrains IDEs, with more clients coming. To try it, update your client, pick Fusion in the model picker, run Standard on everyday work for a few days, then throw one hard problem at Max.