GitHub Copilot Coding Agent Model Picker Now Available for Business and Enterprise

Coding Agent Model Picker Reaches Business and Enterprise Users

GitHub has expanded model selection capabilities for the Copilot coding agent to all paid plan tiers, including Copilot Business and Copilot Enterprise subscribers. When the model picker first launched in December 2025, it was exclusive to Pro and Pro+ plan users. The February 19 update brings that flexibility to organizations and enterprises operating GitHub Copilot at scale.

How the Coding Agent Works

The Copilot coding agent is an asynchronous, autonomous background agent designed to handle delegated programming tasks. When a developer assigns it a task — via github.com, the agents tab, agents panel, GitHub Mobile, or the Raycast launcher — it spins up a cloud-based development environment, works through the task independently, and then opens a pull request for human review.

Available Models

Business and Enterprise users can now select from the following models when assigning tasks to the coding agent:

• Auto (dynamically optimized)
• Claude Opus 4.5
• Claude Opus 4.6
• Claude Sonnet 4.5
• Claude Sonnet 4.6
• GPT-5.1-Codex-Max
• GPT-5.2-Codex
• GPT-5.3-Codex

The "Auto" setting automatically chooses a model based on current availability and task characteristics. When no models have been explicitly enabled by administrators, Claude Sonnet 4.6 serves as the default fallback.

Administrative Considerations

For Copilot Business and Enterprise plans, administrators must ensure that the relevant model policies are enabled in Copilot settings before their teams can access model selection.