GitHub Copilot Fusion is now generally available

What is Copilot Fusion

GitHub Copilot Fusion is a model routing and orchestration system that removes the need for developers to manually select which AI model to use for each coding task. Instead of picking a specific model from a list, users select Copilot Fusion and choose an effort level—Standard for everyday work, High for challenging tasks, or Max for the most demanding problems. The system then determines which models to use and how to combine them for each prompt.

Copilot Fusion combines multi-model orchestration from Project HydraFusion with auto model selection's fine-tuned decision-making. It can execute three different workflows: Single (one model solves the task), Cascade (an efficient model drafts a response, with escalation to a stronger model if needed), and Critique (one model drafts, another reviews, and the original model revises based on feedback).

Performance and cost efficiency

In the Artificial Analysis Coding Agent Index, Copilot Fusion Max achieved a score of 60, placing it within two points of Claude Code with Opus 5.5 Max. More notably, it delivered this performance at approximately half the average API cost per task, used about half as many tokens, and completed tasks in roughly 40% less time. These metrics show that Copilot Fusion can balance capability with efficiency without requiring developers to manually trade off performance against spending.

The system learns from task characteristics and adapts its model selection over time, reducing friction for teams that previously had to remember to switch models between complex and routine work.

Administration and control

Enterprise and organization administrators have granular control over Copilot Fusion. They can specify which models Fusion is permitted to use based on their enterprise and organization model policies. Administrators can independently set each model's policy state and disable models for Fusion without disabling them for other uses.

Copilot Fusion can be set as the default model at any effort level through enterprise managed settings. This allows teams to standardize on automatic routing without requiring individual developers to reconfigure their setup. The system respects existing model availability policies and does not override broader model management decisions.

Billing and limited-time offer

When Copilot Fusion uses multiple models for a single task, billing charges are per model at each model's standard rate. Users on paid Copilot plans will receive a 30% discount on model costs when using Copilot Fusion for a limited time, providing an opportunity to evaluate the cost-benefit of the new system at a reduced rate.

Getting started

Copilot Fusion is available across all Copilot plans in Visual Studio Code, the GitHub Copilot app, GitHub Copilot CLI, and JetBrains IDEs, with support for additional clients coming soon. Developers can access it by updating their client, opening the model picker, selecting Copilot Fusion, and choosing an effort level. Users who participated in earlier research previews with auto, HydraFusion, or HydraFusion Max enabled will see their settings automatically transition to the corresponding Copilot Fusion effort levels.

Source

AI-assisted, verified against the source. This post was drafted with AI from the primary source linked above and checked automatically against it. For anything you depend on, read the original.