On November 10, 2026, GitHub announced that Gemini 4 Carbon was generally available in Copilot for Pro+, Max, Business, and Enterprise plans, with access rolling out gradually. GitHub Docs lists per-million-token rates in U.S. dollars, scheduled to double on December 1. GitHub’s announcement covers access; GitHub’s pricing documentation details the rates.
Which plans and Copilot clients include Carbon?
GitHub lists Copilot Pro+, Max, Business, and Enterprise. It names Visual Studio Code, Visual Studio, Copilot CLI, the GitHub Copilot coding agent and app, github.com, GitHub Mobile on iOS and Android, JetBrains IDEs, Xcode, and Eclipse as places to select the model. The rollout is gradual, so access may not appear for every eligible user at once.
This Copilot integration is separate from Google’s earlier staged-access announcement; NeoTeo covered that earlier announcement.
GitHub positions Gemini 4 Carbon for complex coding, agentic tasks, reasoning, secure code generation, and security-focused work. Those are the uses GitHub highlights for the model.
What Gemini 4 Carbon costs per token
GitHub Docs lists separate rates for input, cached input, and output tokens. The scheduled December 1 increase doubles each listed rate:
| Pricing period | Input per 1 million tokens | Cached input per 1 million tokens | Output per 1 million tokens |
| Before December 1, 2026 | $2.00 | $0.10 | $10.00 |
| Starting December 1, 2026 | $4.00 | $0.20 | $20.00 |
For a simple estimate, one million uncached input tokens plus one million output tokens totals $12.00 at the earlier rates ($2 + $10), or $24.00 at the rates scheduled from December 1 ($4 + $20). That calculation excludes cached input and any plan allowance.
Copilot usage consumes AI credits, and the allowance varies by plan. GitHub says one AI credit equals $0.01 USD; usage beyond an included allowance is charged at the listed per-token rates. The example is therefore a token-rate subtotal, not a prediction of an individual bill.
Check the model’s per-token rates alongside your plan’s AI-credit allowance: the listed rate alone does not determine your bill.
How organizations control access
Business and Enterprise administrators manage model availability through the model policy in Copilot settings. GitHub says new models are enabled by default unless the organization-wide default is off or an administrator disables Carbon.
