Back to rankings

Copilot CLI Discovers Local Ollama Models in /model

GitHub CopilotView original changelog

Starting in version 1.0.94-0, the /model command in GitHub Copilot CLI lists supported models from a running local Ollama instance alongside configured models and Copilot's cloud models. Developers can add a discovered model and use it immediately without restarting the CLI. Choosing a local model does not enable offline mode, which remains opt-in through COPILOT_OFFLINE=true.

Key Takeaways

  • /model now discovers local Ollama models in Copilot CLI version 1.0.94-0 and later.
  • Discovered models appear alongside configured and cloud models in one picker.
  • Models can be added and used mid-session without restarting the CLI.
  • Models must support tool calling and streaming, and Ollama must already be running.
  • Choosing a local model does not enable offline mode, which still requires COPILOT_OFFLINE=true.
  • GitHub previewed intelligent routing with local models as a coming feature.

Local Models in the Model Picker

GitHub added local model discovery to GitHub Copilot CLI. Starting in CLI version 1.0.94-0, the /model command lists supported models from a running local Ollama instance next to the developer's configured models and GitHub Copilot's cloud models.

Adding a Discovered Model

Discovery does not add models automatically. A developer selects a discovered model, reviews its provider and endpoint, and then chooses either "Add and use for this session" or "Add without switching." The model can be used right away, with no CLI restart. If a provider connection fails, the picker explains why.

Requirements

Ollama and the model must already be installed, because the feature neither installs a runtime nor downloads models. Models must support tool calling and streaming. The GitHub Copilot app offers a similar experience, where a supported provider can be added under Settings, then Model providers.

Privacy and Offline Mode

Picking a local model does not enable offline mode or disable GitHub telemetry. Offline mode stays opt-in through COPILOT_OFFLINE=true, and a remote provider can still receive prompts and code context over the network even in offline mode. GitHub also said intelligent routing with local models is coming, with details to follow on the Microsoft Command Line blog.

Copilot CLI Discovers Local Ollama Models in /model | Yet Another Changelog