GitHub makes Copilot Fusion generally available, routing each coding task across AI models
Announced at GitHub Universe on Oct. 28, 2026, Fusion picks and orchestrates models per prompt; GitHub says Fusion Max came within two points of Opus 5.5 Max at about half the cost.
TL;DR
- GitHub made Copilot Fusion generally available on Oct. 28, 2026; it routes and orchestrates which AI models handle each coding task.
- Users pick an effort level (Standard, High or Max); Fusion runs Single, Cascade or Critique workflows across one or more models.
- GitHub says Fusion Max scored 60 on the Artificial Analysis Coding Agent Index, within two points of Claude Code with Opus 5.5 Max, at about half the average API cost per task.
- Fusion is billed per model at standard rates, with a limited-time 30% discount on paid Copilot plans.
GitHub wants developers to stop picking models by hand. On Oct. 28, 2026, at GitHub Universe, the company made Copilot Fusion generally available, a system that decides for each prompt which models to use and how they work together, according to the GitHub changelog.
How it works
- Three effort levels: Fusion Standard for everyday coding, Fusion High for harder tasks and deeper reasoning, Fusion Max for the most demanding work.
- Three workflows: Single (one model solves the task), Cascade (an efficient model drafts, a quality check accepts or escalates) and Critique (one model drafts, another reviews, the first revises).
- Admin control: Fusion only uses models enabled in enterprise and organization policies, and admins can set it as the default.
Performance and pricing claims
GitHub says Copilot Fusion Max scored 60 on the Artificial Analysis Coding Agent Index, within two points of Claude Code with Opus 5.5 Max, at about half the average API cost per task, using about half as many tokens and about 40% less time per task. Users are billed for each model Fusion uses at standard rates; paid Copilot plans get a limited-time 30% discount on model costs. Fusion is available on all Copilot plans in VS Code, the GitHub Copilot app, Copilot CLI and JetBrains IDEs.
The wider Universe package
In his keynote post, Jay Parikh, Executive Vice President of CoreAI at Microsoft, also introduced Copilot projects, session continuity across devices, Copilot code review Max and access to all Copilot features for GitHub Enterprise customers without a separate subscription, billed by AI credits used. GitHub says it has added 4 million CPU cores, 160 PB of high-speed storage and 50 Tb/s of network capacity so far in 2026. Unite.AI notes Fusion follows the Project HydraFusion research preview introduced on Sept. 4, 2026.
What it means for Web3 teams – Analysis
Automatic model routing can cut AI coding spend for small smart-contract and dApp teams that default to the most expensive model. Benchmarks are GitHub's own; audit-critical code still needs human review. This is not investment advice.
FAQ
What is GitHub Copilot Fusion?
A system, generally available since Oct. 28, 2026, that routes and orchestrates which AI models Copilot uses for each task instead of the user picking one model.
How is Copilot Fusion billed?
Per model used, at each model's standard rate; paid Copilot plans get a limited-time 30% discount on model costs.
Where is it available?
On all Copilot plans in Visual Studio Code, the GitHub Copilot app, GitHub Copilot CLI and JetBrains IDEs.
How does it perform?
GitHub says Fusion Max scored 60 on the Artificial Analysis Coding Agent Index, within two points of Claude Code with Opus 5.5 Max, at about half the average API cost per task.
Sources
News summary based on the sources above. Not investment, legal or tax advice. Editorial policy & corrections · Report an error
Building in Web3? Turn the news into trust.
Positioning, KOL programs, LinkedIn systems and claim-safe content for Web3 and fintech teams.
Related
Fake Claude Code update in Cursor's marketplace tried to plant an SSH backdoor key, up to 5,000 users hit
A tampered Anthropic.claude-code 2.1.313 build sat in the Cursor extension marketplace for about 11 hours on Oct. 29, 2026. Cursor says uninstalling it may not be enough.
Intel shares plunge as Q4 guidance of $15–16 billion misses forecasts despite 78% data center jump
Intel beat Q3 2026 estimates with $16.5 billion in revenue, but a weak fourth-quarter outlook, a PC slowdown and a deadly fab explosion sent the stock down about 30% on Oct. 30, 2026.
SK hynix posts first 100-trillion-won quarter as AI memory demand lifts profit 632%
Q3 2026 revenue hit 104.6 trillion won and operating profit 83.3 trillion won, both above analyst estimates, and the memory maker plans about 100 trillion won in shareholder returns.