GitHub Copilot Fusion Handles Model Selection and Orchestration for Coding Tasks

GitHub has debuted GitHub Copilot Fusion, a system that handles model selection and orchestration for coding tasks so developers do not have to. The tool combines orchestration from Project HydraFusion with Copilot Auto Model Selection and is generally available now. GitHub announced it Wednesday at GitHub Universe in San Francisco.

What Copilot Fusion Does

For tasks involving reasoning, code generation, debugging, or tool use, developers can delegate decisions about which models to use and how they work together. GitHub says developers can choose Copilot Fusion and let Copilot intelligently route and orchestrate which models it uses. Different tasks benefit from different models based on capabilities and cost, so the system can switch models mid-workflow when a step calls for it.

After analyzing what a task requires and choosing a model, Copilot Fusion orchestrates the next steps. That may be one model doing everything, a cascading workflow where a quality check evaluates one model's response and then accepts or escalates it, or a critique workflow where one model reviews another's work. GitHub says this could reduce mental load for developers who may choose inefficient models when their attention drifts.

Performance and Cost

GitHub says it has improved performance since launching Project HydraFusion as a research preview in September 2026. On the Artificial Analysis Coding Agent Index, Copilot Fusion Max scored 60, within two points of Claude Code with Opus 5.5 Max. GitHub says Fusion Max delivered that result at about half the average API cost per task, using about half as many tokens and taking about 40% less time per task.

GitHub describes that as comparable benchmark performance. In a statement supplied to The New Stack, Hari Pokhrel, principal software engineer at AT&T, credits Copilot Fusion Standard with improving token efficiency at AT&T, reducing unnecessary spending, and preserving capacity for higher-value work.

Controls and Availability

Developers choose the effort level while Copilot Fusion handles model selection and orchestration. Fusion Standard is for everyday coding, Fusion High is for more challenging tasks and deeper reasoning, and Fusion Max is for the most demanding and complex work. Enterprise administrators can control which models Copilot Fusion can choose from, disabling certain models for Copilot Fusion without disabling them for the entire developer team.

Administrators can also adjust the policy state for each model independently and set whether Copilot Fusion is or is not the default model for Copilot users. GitHub says the product is part of its vision to make GitHub the most comprehensive software delivery system in the world, with developers delegating model choices while retaining responsibility for the direction and outcome of the work.

Model Routing Trend

Copilot Fusion was released alongside the GitHub Copilot app, Copilot Projects, and Copilot Automations. GitHub says deferring model selection to its system can give developers more time and focus for other work. This month, Ramp described moving model selection into Open SWE, its open-source coding agent, and Replit made intelligent model routing the default across every account.

Sources: The New Stack