AI BriefingGitHub Copilot Fusion generally available with 60-point benchmark score
·2026.10.28 09:00
Key point
The new system routes and orchestrates models automatically, achieving a score of 60 on the Artificial Analysis Coding Agent Index while using about half the tokens and 40% less time per task.
Details
- GitHub Copilot Fusion is now generally available, replacing manual model selection with an intelligent system that routes and orchestrates models for each task.
- Copilot Fusion Max scored 60 on the Artificial Analysis Coding Agent Index, within two points of Claude Code with Opus 5.5 Max, while using about half the average API cost per task.
- The system uses approximately half as many tokens and takes about 40% less time per task compared to the benchmark comparison.
- Users select one of three effort levels: Fusion Standard for everyday coding, Fusion High for challenging tasks, and Fusion Max for the most demanding complex reasoning.
- Administrators can control which models Fusion uses via enterprise model policies and set it as the default model for users.
- Billed at each model's standard rate, with a limited-time 30% discount on model costs for users on paid Copilot plans.
- Available on all Copilot plans in Visual Studio Code, the GitHub Copilot app, GitHub Copilot CLI, and JetBrains IDEs.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
Yogiyo Product Engineer Pilot: Context and Verification, Not Code Generation, Are the Bottlenecks
Microsoft Makes GitHub's Multi-Model Copilot Harness Available via Microsoft FoundryGitHub Copilot Automations now learn from outcomes to self-improve prompts