Noesion Radarnoesion.ai
Release2026-10-28

GitHub Copilot Fusion now supports intelligent routing for local models. It can choose between on-device and cloud models in Copilot CLI, the Copilot app and VS Code. The feature is in experimental preview and must be opted into. Windows users can download and configure MAI Code 1.1 Flash through Windows ML. Users on Windows, macOS and Linux can bring their own local models via an OpenAI-compatible server. Supported models include MAI Code 1.1 Flash and Qwen3.8 Flash Next. Local model auto routing must be enabled in settings. Selecting Standard in Copilot Fusion allows automatic model selection. Local inference does not make sessions offline. Routing decisions are still made in the cloud. Local models may have different safety protections than cloud models. The Copilot app’s intelligent routing currently supports managed Windows ML models only.

What changed

Local models had to be selected manually for each request.

What it unlocks

Letting GitHub Copilot Fusion automatically choose between local and cloud models for each request.

What you need to act on it

  • experimental mode enabled
  • supported device and OS
  • sufficient local compute

Sources

Send this to someone who needs it

Shares the story and its sources. Nothing about you.

What does this mean for your job?

This is the story as everyone gets it. Once a week we send you the version written for your role — what changed, why it matters for the work you actually do, and one thing to try. Free while we tune it.