GitHub Copilot Shifts to Automatic Model Enablement for Enterprise Users
Key Takeaways
- ▸GitHub Copilot models will automatically enable for Copilot Business and Enterprise users upon general availability, eliminating the need for manual admin configuration starting August 26, 2026
- ▸Organizations have a 28-day transition period to review settings and opt-out of automatic enablement if they prefer to approve each model manually
- ▸Prior explicit model configurations are preserved; only previously unconfigured models will follow the new default policy
Summary
GitHub is introducing a default enablement policy for generally available AI models on its Copilot Business and Enterprise plans, automating the rollout of new models to users instead of requiring manual administrator configuration. Starting August 26, 2026, new models will be turned on by default for these plans, though organizations can opt-out if they prefer stricter governance controls. The change follows a 28-day grace period that allows enterprises to review their model settings and decide whether to embrace the new default or disable it for more granular control.
The policy preserves backward compatibility by maintaining all previously configured explicit choices—organizations that have deliberately enabled or disabled specific models will see those settings honored. Certain models, including open-weight options like DeepSeek and proprietary models not covered by GitHub's data retention agreements, are excluded from automatic enablement regardless of policy settings. This approach aims to streamline AI model adoption in enterprise productivity workflows while giving organizations the flexibility to maintain tighter governance if needed.
- Open-weight models and certain proprietary models not covered by GitHub's data agreements are excluded from automatic enablement
