Newsbite analysis · Programming, Tools
GitHub Blog
Discover local models in GitHub Copilot CLI
The latest GitHub Copilot CLI version allows users to discover supported local models from a running Ollama instance without disrupting their workflow. Users can select models, confirm their addition for use in the current session, and avoid needing to restart the CLI. This update enhances the flexibility of model usage while maintaining connectivity to remote providers for prompts and context even in offline mode.
Act nowConfidence: HighImmediate horizonStory: First report
EvidenceOfficial announcementNo quotes
Published 7 Oct 2026, 18:45 UTC2 min read
Lead image — GitHub Blog
Brief
NewsBite reading:Select Local Models in GitHub Copilot CLI
TL;DR
- Introduces local model discovery without interrupting workflow
- Supports models running on a local Ollama instance
- Users can review model providers and endpoints before addition
- Maintains network connections for remote providers even in offline mode
- Incorporates intelligent routing for better model management
Interesting facts
- Discovery does not automatically add models, which gives users control over their selection.
- Models must support tool calling and streaming for compatibility.
- The update mimics the model selection process available in the GitHub Copilot app.