Starting with CLI version 1.0.94-0, the /model command can list supported models from a running local Ollama instance alongside configured and GitHub Copilot cloud models.
Discovery does not automatically add a model. A user selects a discovered model, reviews its provider and endpoint, and confirms either “Add and use for this session” or “Add without switching.” The selected model can be used without restarting the CLI. Ollama and the model must already be installed, and eligible models must support tool calling and streaming. Connection failures are shown in the model picker with an explanation.
GitHub also states that selecting a local model does not enable offline mode or disable telemetry. Offline mode remains an explicit setting through COPILOT_OFFLINE=true; a remote provider may still receive prompts and code context over the network, including in offline mode. The changelog announces intelligent routing with local models but does not state its availability, pricing, supported model list, or the conditions under which routing will be used.