GitHub Copilot Fusion is now generally available. It routes and orchestrates models for each task automatically. Users pick an effort level: Standard, High or Max. It runs three workflows: Single, Cascade and Critique. Admins can set which models Fusion uses and make it the default. Billing is per model used at standard rates. A 30% discount on model costs applies for a limited time. Fusion is available in VS Code, the Copilot app, CLI and JetBrains IDEs. Previous auto or HydraFusion selections transition to Fusion levels automatically.
What changed
Developers had to pick a single model for each task and switch manually.
What it unlocks
Letting Copilot pick and orchestrate models per task without manual switching.
- 60 on Artificial Analysis Coding Agent Index
- half the average API cost per task
- half as many tokens
- 40% less time per task
What you need to act on it
- a paid Copilot plan
- client update
Sources