On November 12, 2026, Google DeepMind launched Gemini 4 Hydrogen, a new member of the Gemini 4 family built for speed and execution efficiency — a fast model for everyday coding and agentic workflows. Unlike Gemini 4 Carbon and Argon, which focus on deep, long-horizon reasoning, Hydrogen is built to do a different job: get work done fast and reliably, at a lower cost.
According to the official blog, Hydrogen scores 51 on the Artificial Analysis Intelligence Index at its default high thinking level, ahead of Argon's 50; at the new xhigh level it rises to 54, just one point behind Claude Sonnet 5.5 and GPT-6.1 Sol. Pricing is flat across the 1-million-token context window: $1.50 per million input tokens and $7.50 per million output tokens. It runs 2.5 times faster than the previous 3.8 Flash, jumps from 18.2% to 57% on Terminal-Bench 4.1, and cuts fabricated claims by more than fivefold.
Hydrogen's real selling point is not working alone but working as part of a multi-model team. DeepMind also launched Orchestrator in Google Antigravity, which switches automatically between Carbon and Hydrogen within a single session: heavy architecture decisions go to Carbon, while writing code, running tests and fixing bugs go to Hydrogen. On Terminal-Bench 4.0, Orchestrator reaches task success rates close to a pure Carbon setup while using 24% less quota on average and finishing 19% faster. When several agents run in parallel, Carbon acts as the coordinator while Hydrogen powers the worker and verifier agents.
Starting November 12, Hydrogen is rolling out to developers, enterprises and everyday users: developers through Google AI Studio and the Gemini API, Pro and Ultra subscribers through Google Antigravity, enterprises through Gemini for business, and consumers directly in the Gemini app and Google Search. For everyday users the change is more tangible: Search understands real intent better, answers are more visual, and multi-turn conversations in the Gemini app follow instructions more reliably.
Hydrogen is no all-rounder: hard tasks that need deep architectural judgment and long reasoning remain Carbon and Argon's territory — Hydrogen wins on execution. Google has not yet announced Hydrogen's API rate limits or regional rollout schedule, so the real-world experience will depend on how the rollout proceeds. Developers who want to try it can switch models in AI Studio, start with everyday coding tasks, and then decide whether to put it into a production workflow.
