Copilot Fusion Picks the Models. Check the Whole Job’s Bill.
Maestro Brief · Published by Maestro Mojo
2026-10-29
Maestro’s take. Try Copilot Fusion as a routing experiment, not a guaranteed saving. Let it pick the models. Keep your own acceptance test and record the full cost of the job.
TL;DR
GitHub made Copilot Fusion generally available on October 28. Users choose Standard, High or Max effort instead of a specific model.
The Fusion documentation describes a system that can use one model, escalate a draft to a stronger model, or add a separate critique-and-revision step. The unit to evaluate is that whole workflow.
GitHub bills each model used at its standard rate. The documentation gives paid Copilot users a 30% discount on Fusion model costs through November 30, 2026.
Why Maestro users care
Routing can save the effort of matching every task to a model. Standard selects one model per task. High and Max can add model calls for review or escalation. Neither proves the final job will be cheaper or better.
One limitation deserves attention before a trial: if Fusion discards a draft, it does not automatically undo file edits that draft already made. Review the complete workspace diff before committing, not just the final answer.
Keep the approved model pool deliberate. GitHub’s enterprise controls let administrators configure Fusion’s available models. Those settings still respect the user’s other model policies.
Maestro’s rule: evaluate the router with the model pool you will actually allow. A result from a different pool does not answer your team’s question.
One thing to try
Run a small set of routine fixes through Fusion Standard and your current setup. Keep the tests and review bar the same, and inspect the full diff before accepting a fix. Record accepted fixes, elapsed time, rework and the complete model bill. Compare both today’s discounted cost and the cost without the promotion.
Published October 29, 2026 · Brief and Maestro commentary · Tags: AI coding, costs, workflow
Original sources: GitHub’s launch and billing terms, Fusion’s execution patterns, and enterprise model controls.