Google has introduced Gemini 4 Hydrogen as a model for everyday coding and agent workflows, alongside a system that switches models as work develops. The announcement, dated November 12 and updated November 13, describes a rollout across its developer, business and consumer services.
Hydrogen reaches the Gemini API, AI Studio, the Gemini app and Search. Free Gemini app users receive access through automatic model selection when appropriate. Google lists API pricing of $1.50 per million input tokens and $7.50 per million output tokens across a one-million-token context window.
In Antigravity, paid AI Pro and Ultra subscribers gain Hydrogen, while Orchestrator can move between models within a session. Google’s example gives Carbon the planning role, Hydrogen the implementation and testing, and Carbon another turn if execution encounters difficulties. Plan restrictions matter: Pro gets Orchestrator with Argon; Ultra also gets the Carbon option.
For development teams, the change makes model selection part of execution rather than a single decision at the start. Google reports efficiency and benchmark gains, but these are vendor-reported results, not independent evidence that a particular codebase will see the same improvement. Teams should evaluate the workflow against their own tests, review requirements and usage limits.
