Intelligent local model routing is coming to GitHub Copilot

Name: Intelligent local model routing is coming to GitHub Copilot
Uploaded: 2026-10-07T18:10:22+00:00
Description: GitHub explains an upcoming GitHub Copilot feature that intelligently routes some requests to local, on-device models (including Ollama and Microsoft...

Yesterday by GitHub

GitHub shares an upcoming GitHub Copilot capability: intelligent model routing that extends model orchestration from the cloud down to the local machine.

Overview

The video explains how GitHub Copilot will be able to automatically discover installed local models from providers such as Ollama and Microsoft Foundry Local, without manual setup.

Key points covered:

Automatic local model discovery
- Copilot can detect locally installed models (for example via Ollama and Microsoft Foundry Local) without requiring manual configuration.
Intelligent routing to local compute
- Copilot will be able to auto-route certain “simple tasks” to an on-device model.
- This is positioned as using free local compute for those tasks.
Zero AI credits for local workloads
- Some workflows (including codebase explanations and long-running/background automations) can be offloaded to local models for zero AI credits.
Offline usage and data protection
- Local routing is presented as enabling Copilot workflows to run online and completely offline.
- Keeping execution on-device is also framed as a way to better protect sensitive data.
Availability
- The video includes a short segment on availability and what’s coming next.

Learn more: https://gh.io/localdevelopment