The Pulse
GitHub Makes Copilot Fusion Generally Available Across Plans
GitHub has made Copilot Fusion generally available across Copilot plans, letting the system choose and orchestrate models for coding tasks. The feature is available in four clients, and GitHub says paid-plan users receive a 30% discount on
AI.info Team · Written with AI · fact-checked ·
GitHub says developers can stop choosing a model for every coding task; its new Copilot Fusion will route work among models instead. But the system does not make those model calls free: users pay each model’s standard rate, and more involved tasks can consume additional AI credits. GitHub announced Fusion’s general availability on October 28, 2026, at GitHub Universe.
Select Fusion in the model picker
Fusion combines Project HydraFusion’s multi-model orchestration with Copilot’s automatic model selection. A developer selects Fusion and an effort level rather than naming an individual model; Copilot decides which models to use and how they work together for each prompt. GitHub offers Standard for everyday coding, High for harder tasks and deeper reasoning, and Max for demanding work.
The system can take three routes. In Single, one model handles the task; in Cascade, an efficient model drafts a response and a quality check can send it to a stronger model; in Critique, one model drafts, another reviews, and the first revises. The route can change from prompt to prompt within a session, according to GitHub’s product documentation.
Efficiency claims come with a bill
GitHub says Fusion Max scored 60 in the Artificial Analysis Coding Agent Index, within two points of Claude Code with Opus 5.5 Max. The company says Fusion reached that result at about half the average API cost per task, using about half as many tokens and taking about 40% less time. Those are GitHub’s reported benchmark results, not a guarantee that every team or workload will see the same savings.
When using Copilot Fusion, users are billed for each model it uses at that model’s standard rate, with no separate charge for Fusion. High and Max can use more than one model for a task, so a task at either level may use more AI credits than one handled at Standard. GitHub’s documentation says users on paid Copilot plans receive a 30% discount on model costs through November 30, 2026.
Administrators set the boundaries
Organizations retain control over the model pool: Fusion can use only models enabled under the organization’s and enterprise’s policies. Administrators can disable a model for Fusion without removing access to that model elsewhere, and can set Fusion as the default for Copilot users through enterprise-managed settings. GitHub also says the available models can change as it evaluates them and adds new ones.
That control matters because Fusion’s promise is to move model selection out of the developer’s hands, not out of the organization’s. Hari Pokhrel, principal software engineer at AT&T, described the intended benefit:
“GitHub Copilot Fusion Standard has improved token efficiency across AT&T by automatically selecting the right model for the task, reducing unnecessary spend and preserving capacity for higher-value work. Previously, developers would often choose a more powerful model for a complex task and forget to switch back for everyday activities. Copilot Fusion eliminates that friction, optimizing model selection without requiring developers to manage it themselves. We’ve also seen its decision-making improve over time, giving developers greater confidence to focus on their work while Copilot handles model selection.”
— Hari Pokhrel, Principal Software Engineer, AT&T
From HydraFusion preview to four clients
Fusion is available on all Copilot plans in Visual Studio Code, the GitHub Copilot app, GitHub Copilot CLI, and JetBrains IDEs. GitHub’s October 28 Octoverse report says the HydraFusion research preview routed almost 60,000 sessions in the CLI and Copilot app. That orchestration work is now generally available as Fusion. Users who had enabled the earlier auto, HydraFusion, or HydraFusion Max options will move to corresponding Fusion effort levels.
GitHub says support for additional clients is coming. For now, developers can choose Fusion in the model picker in the four supported clients; paid-plan users receive a 30% discount on model costs through November 30, 2026.