GitHub Copilot Fusion is now available on every Copilot plan

GitHub made Copilot Fusion generally available on October 28, according to its changelog post. Fusion is a model choice that picks which models to use for each task, and how they work together, so you don't have to choose one yourself.

The pitch is cost. GitHub says Fusion Max scored 60 on the Artificial Analysis Coding Agent Index, within two points of Claude Code running Opus 5.5 Max, at about half the average API cost per task.

How Copilot Fusion works

You select Fusion in place of a single model and pick an effort level. Standard uses versatile models for everyday coding. High uses more capable models for harder tasks and deeper reasoning. Max uses the strongest models for the most demanding work.

Behind that choice, Fusion runs one of three patterns:

• Single: one model does the task.
• Cascade: a cheaper model drafts an answer, and a quality check either accepts it or hands the task to a stronger model.
• Critique: one model drafts, a second reviews, and the first revises.

Fusion combines the multi-model orchestration from Project HydraFusion with Copilot's automatic model selection. Anyone who used the research preview as Auto, HydraFusion or HydraFusion Max is moved to the matching Fusion effort level.

What GitHub says about speed and cost

GitHub's numbers come from its own run of the Artificial Analysis Coding Agent Index. It says Fusion Max used about half the tokens and about 40% less time per task than Claude Code with Opus 5.5 Max, and cost about half as much in API fees. That is GitHub's claim about its own product, and the changelog doesn't say how many tasks were run. Check Artificial Analysis for the index itself.

Where it is available and what it costs

Fusion is on every Copilot plan, in Visual Studio Code, the GitHub Copilot app, GitHub Copilot CLI and JetBrains IDEs. GitHub says more clients are coming.

You pay for each model Fusion uses, at that model's normal rate. For a limited time, people on paid Copilot plans get 30% off model costs when they use Fusion. GitHub doesn't say when the discount ends. The post also doesn't list which models Fusion routes between or give premium request multipliers or usage limits, so look in the Copilot Fusion docs before relying on it for a large job.

For organizations, Fusion only uses models that the enterprise or organization has enabled. An enterprise admin can switch off specific models for Fusion alone, and can make Fusion the default model at any effort level through managed settings.

Fusion landed alongside the rest of GitHub Universe, where GitHub also released the Copilot app and said GitHub Enterprise now includes Copilot. For how Copilot's plans compare with a rival, see our Cursor vs GitHub Copilot comparison.

What to watch

The two open questions are which models sit behind each effort level, and what happens to the price when the 30% discount ends. Independent runs of the Coding Agent Index would also show whether the score holds up outside GitHub's own test.