GitHub brings Gemini 4 Hydrogen to Copilot across five paid plans

GitHub made Gemini 4 Hydrogen generally available in Copilot on November 13, 2026, giving developers another model for coding work across editors, command-line tools and coding agents. The release covers five paid plans, including Copilot Pro, but access is rolling out gradually. Usage is metered at provider list pricing.

One model choice across editors and agents

Eligible subscriptions are Copilot Pro, Pro+, Max, Business and Enterprise. Users choose Hydrogen through Copilot’s model picker rather than a separate application. GitHub lists support across desktop development environments, its own coding tools, the web and mobile, as well as the Copilot software development kit.

• Editors and development environments: Visual Studio Code, Visual Studio, JetBrains IDEs, Xcode and Eclipse.
• Copilot tools: Copilot CLI, the command-line interface; GitHub Copilot coding agent; and the GitHub Copilot app.
• Other supported interfaces: github.com, GitHub Mobile on iOS and Android, and Copilot SDK.

That is the announced reach of the release, not a promise that every eligible account already has the option. GitHub says users who cannot see Hydrogen should check again as the rollout progresses. General availability and immediate visibility in the model picker are therefore different things for this release.

The pitch is faster execution, not just deeper reasoning

GitHub describes Hydrogen as Google’s latest Flash model, designed for fast, efficient agentic coding: work in which an AI uses tools and carries out multiple steps. It also cites improved instruction following over Gemini 3.8 Flash. Those are GitHub’s descriptions of the model’s intended role and improvement.

Google introduced Hydrogen on November 12. Its launch announcement draws a division of labor within the Gemini 4 family: Carbon and Argon handle deep architectural judgment and difficult reasoning, while Hydrogen emphasizes rapid execution. Google says Hydrogen can work alone or handle writing code, running tests and checking changes alongside those models.

For teams moving from Gemini 3.8 Flash, Google claims Hydrogen finishes tasks 2.5 times faster at its default high thinking setting. It also reports a large gain on Terminal-Bench 4.1. These are Google-published model results, not measurements of how Hydrogen performs inside GitHub Copilot.

Metered use, with access enabled by default

GitHub says Hydrogen use in Copilot is billed at provider list pricing under usage-based billing. Google lists the model at $1.50 per million input tokens and $7.50 per million output tokens, with flat rates across its 1-million-token context window. Tokens are the units of input and output used for that pricing.

Workplace access has a separate control. Copilot Business and Enterprise administrators can manage Hydrogen through the model policy in Copilot settings. Under GitHub’s default model-enablement behavior, new models are automatically enabled unless an administrator has switched off the global default or explicitly disabled the particular model.

For those organizations, the practical decision is not necessarily whether to turn Hydrogen on. With default enablement left in place, administrators who want to withhold access need to disable the model or change that global setting.