Google Launches Gemini 4 Hydrogen: Its Fast, Cheap Model for Coding and AI Agents — and It's in Search for Everyone

Photo: Hippopx (public domain). Illustrative image.
What happened. Google introduced Gemini 4 Hydrogen on 12 November 2026, the third model in its Gemini 4 family and the first meant for everyone. Where Carbon and Argon are the frontier models, built for deep, long-horizon reasoning, Hydrogen is built for rapid execution: writing code, running tests, verifying changes, doing research, and carrying multi-step tasks to completion. Google calls it its “daily driver for agentic workflows”. The launch was announced on Google’s blog by Koray Kavukcuoglu, senior vice president at Google DeepMind and Google’s chief AI architect.
What it costs. Hydrogen is priced flat across its entire 1 million-token context window at $1.50 per million input tokens and $7.50 per million output tokens. There is a free tier in AI Studio. The flat pricing matters for agents: long-running coding and research jobs chew through context, and a model whose price does not rise as the context fills up is easier to budget for than one that charges more for the long tail.
What Google claims it can do. At its default high thinking level, Google says Hydrogen scores 51 on the Artificial Analysis Intelligence Index, edging out Argon at 50, and reaches 54 at its highest thinking setting, xhigh. There are four configurable thinking levels: low, medium, high and xhigh. Google also says Hydrogen finishes tasks 2.5 times faster than Gemini 3.8 Flash, and jumps from 18.2% to 57% on Terminal-Bench 4.1, a benchmark for agentic coding. All of these figures are Google’s own; independent verification will come as third-party evaluations publish. The honest reading of the scorecard is that Hydrogen is roughly at parity with Argon on one index, not clearly ahead of it — its case is speed and price, not raw capability. Our comparison of the Gemini 4 family against Claude and Muse shows how close the frontier race already is.
Where you can use it. Starting on 12 November, Hydrogen is rolling out globally across Google’s developer, enterprise and consumer surfaces: the Gemini API and AI Studio on all tiers, Gemini for business, the Gemini app, Android Studio, Google Sheets, and Google Antigravity for paid Google AI Pro and Ultra subscribers — where Google also shipped a new Orchestrator for multi-agent work. And it is in Google Search for all users, globally, which Google calls its biggest model upgrade to Search yet. That is the notable part: the two earlier Gemini 4 models went into Search only for paying subscribers, while Hydrogen goes to everyone. Google’s Search post, by product vice president Robby Stein, says Hydrogen brings more helpful, visual and intuitive responses, with improvements to intent, links, visual answers and language — but gives no numbers for any of it, and says nothing about advertising or what happens to clicks to websites.
Why it matters. Hydrogen is Google’s answer to a shift that has been building all year: the model that wins the agent era may not be the smartest one, but the one that is fast and cheap enough to run all day. It follows Google’s Gemini 4 Carbon launch by a day, and arrives the same week as Google’s Gemini Decision 1, which splits cheap yes/no scoring off from full text generation entirely. The pattern is the same in both: the frontier model does the hard thinking, and a smaller, faster, cheaper model does the routine execution. For developers building agents, the question is no longer which single model to use, but which model for which step — and Hydrogen is Google’s bid to own the execution step.
Sources
- Google Blog — “Introducing Gemini 4 Hydrogen for Agentic Workflows” (12 November 2026): https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-hydrogen/?ref=testingcatalog.com
- AI Weekly — “Gemini 4 Hydrogen Ships at $1.50/$7.50 as Agent Daily Driver” (12 November 2026): https://aiweekly.co/alerts/gemini-4-hydrogen-ships-at-150750-as-agent-daily-driver
- DEV Community — “Gemini 4 Hydrogen: Google’s Fast Execution Model Beats Argon on One Benchmark — Here’s the Catch” (13 November 2026): https://dev.to/tng_thnhnhn_c933ee88b/gemini-4-hydrogen-googles-fast-execution-model-beats-argon-on-one-benchmark-heres-the-catch-2ch6
- PPC Land — “Google Search gains Gemini 4 Hydrogen for all users, not just subscribers” (12 November 2026): https://ppc.land/google-search-gains-gemini-4-hydrogen-for-all-users-not-just-subscribers/
- FreshSynapse — “Google’s Gemini 4 Carbon Is Out of the Lab and Into the Gemini App — for Ultra Subscribers”: https://freshsynapse.com/articles/google-gemini-4-carbon-available-2026-11-11/
- FreshSynapse — “Google’s New AI Doesn’t Write Answers — It Decides: Gemini Decision 1 Scores 1,000 Questions in One Pass”: https://freshsynapse.com/articles/google-gemini-decision-1-decisions-api-2026-11-12/
- FreshSynapse — “Claude Fable 5.2 vs Gemini 4 Carbon vs Muse Ember: Which AI Model for Which Job?”: https://freshsynapse.com/articles/claude-fable-vs-gemini-carbon-vs-muse-ember-2026-11-11/