GitHub Copilot Fusion adds local model routing in experimental preview
Overview
GitHub ships intelligent routing between on-device and cloud models in GitHub Copilot Fusion as an opt-in experimental preview, available in GitHub Copilot CLI, the GitHub Copilot app, and Visual Studio Code.
The routing lets users run requests on their device's compute without switching models manually for each request. GitHub says routing decisions are still made in the cloud, so local inference does not make a session offline.
On supported Windows devices, users can download and configure MAI Code 1.1 Flash through the Windows ML provider. Models added through bring your own key (BYOK) configuration are not automatically eligible for Copilot Fusion, and the routing currently supports managed Windows ML models only.
Written by AI from the articles below · updated Oct 28, 1:03 PM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
- GitHub Copilot ChangelogOfficialPickGitHub Copilot Fusion adds local model routing in experimental preview
AIGitHub Copilot Fusion can now choose between supported on-device models and cloud models in GitHub Copilot CLI, the GitHub Copilot app, and Visual Studio Code. The opt-in experimental preview is set up differently per tool, with managed MAI Code 1.1 Flash download on supported Windows devices. Routing decisions are still made in the cloud, so local inference does not make a session offline.
Heat Trend
Not enough continuous observations to show a trend yet.