GitHub Universe kicked off today with a pile of Copilot news. Here’s what matters to me.
The highlights
- Copilot Fusion is GA. Pick it instead of a model and it routes and orchestrates models per task. 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 API cost and tokens.
- Local model routing (experimental). Fusion can choose between on-device and cloud models in Copilot CLI, the Copilot app and VS Code. The managed download is Windows-only, but CLI routing works on Windows, macOS and Linux with MAI Code 1.1 Flash or Qwen3.8 Flash Next behind a local OpenAI-compatible server.
- GitHub Agentic Workflows is GA. Write automation in Markdown and it compiles to standard Actions YAML, with Copilot or your own coding agent.
- Copilot app Projects (public preview) give multi-repo work one home with its own agents and automations.
- Copilot comes with GitHub Enterprise for enterprises working with an account team, with usage metered against the budget.
The Register adds that GitHub saw 7.4 billion commits in September, five times a year ago, and is adding four million CPU cores and 160PB of storage to keep up.
My take
Local routing on Linux is the one I’ll try first: point the CLI at a local server and let it decide when to call the cloud. Markdown-to-Actions workflows are the same “agents as files” idea everybody’s landing on.
God Bless, and May the Source Be with You!
References & Further Reading
- Route and orchestrate models with GitHub Copilot Fusion (GitHub Changelog)
- GitHub Copilot Fusion now supports selection of local models (GitHub Changelog)
- GitHub Agentic Workflows generally available (GitHub Changelog)
- GitHub Copilot app, Universe releases (GitHub Changelog)
- GitHub Enterprise now includes access to GitHub Copilot (GitHub Changelog)
- GitHub powering up four million extra cores (The Register)
