Allison reports on the new model picker for GitHub Copilot coding agent, available to Copilot Pro and Pro+ subscribers, enabling flexible selection between multiple advanced AI models.

Model Picker Now Available for Copilot Coding Agent (Pro and Pro+ Subscribers)

GitHub has introduced a model picker feature for the Copilot coding agent, now accessible to Copilot Pro and Pro+ subscribers. This asynchronous, autonomous agent can be used from the agents tab, agents panel, issue assignments on github.com, or through the Raycast extension.

Key Features

  • Model Selection: Users can now select between multiple advanced AI models while initiating a task:
    • Auto (system-optimized selection)
    • Anthropic Claude Opus 4.5
    • Anthropic Claude Sonnet 4.5
    • OpenAI GPT-5.1-Codex-Max
  • Auto Option: If ‘Auto’ is selected, Copilot will optimize for speed and performance by dynamically choosing a model based on availability.
  • Eligibility: Currently available for Copilot Pro and Pro+ subscribers. Expansion to Copilot Business and Enterprise subscribers is planned.
  • Integrations: Model picker will soon be available in more integrations, such as:

How to Use the Model Picker

  • When starting a task via Copilot coding agent-enabled interfaces, look for the model picker dropdown.
  • Choose one of the supported models or leave on Auto for Copilot to decide.

Documentation

For setup and usage details, refer to official documentation about model selection in Copilot coding agent.

This update empowers developers using Copilot with greater flexibility over the AI powering their code-related tasks, allowing customization for speed or advanced reasoning per workflow requirements.

This post appeared first on “The GitHub Blog”. Read the entire article here