Google put a third model in the Gemini 4 family on November 12, 2026, and this one is for everyday work. Gemini 4 Hydrogen is built for speed on everyday coding and agentic workflows, and it shipped to every Google surface the same day: all tiers of Google AI Studio and the Gemini API, Gemini for business, the Gemini app, Google Search, and Google Antigravity for paid AI Pro and Ultra subscribers. One line on Google's pricing page is worth reading before anything else, so we checked it first. The Hydrogen row reads Free of charge for both input and output on the Gemini API's free tier. Two days earlier, when Carbon opened up, its row read Not available.
What Google released
Hydrogen is the third model in the Gemini 4 family. Google's announcement describes Carbon and Argon as the frontier models built for deep, long-horizon reasoning on the hardest problems, and Hydrogen as built for rapid execution across diverse agentic workflows. The model card says Hydrogen was adapted from Carbon, with a 1 million-token context window and text, image, audio, video, and PDF as input.
On benchmarks, Google's announcement cites a 51 score on the Artificial Analysis Intelligence Index, against 50 for Argon at Hydrogen's default thinking level, and says Hydrogen often completes tasks faster than other models in its class. These are benchmarks Google chose for its own announcement.
The model has four thinking levels: low, medium, high (the default), and a new xhigh level for harder problems. Google also upgraded Search with Hydrogen, saying it understands multi-part and typo-filled queries better, gives more visual responses, and pulls a wider range of links.
Why a beginner might care
Hydrogen has a free Gemini API tier for input and output. The same pricing table says that on the free tier, your prompts and responses are used to improve Google's products. The paid tier doesn't use your data that way.
In the Gemini app, Google says Hydrogen follows your tone, format, and specific rules more closely, so email drafts, study guides, rewrites, and documents come back ready to use with fewer corrections. It's also better at calling the right tools for a job.
What we checked on launch day
The app picker lists Hydrogen, and it runs. We opened the Gemini app on November 12 with one paid Google work account. The control next to the prompt box opened a picker listing Everyday help (4 Hydrogen), Complex problems (4 Hydrogen Thinking), and Detailed research (3.1 Pro). We selected Everyday help, ran a rewrite prompt, and the response arrived with a footer that reads 4 Hydrogen.
The pricing page has a free tier for it. Google's API pricing page lists gemini-4-hydrogen with input and output free of charge on the free tier. The paid tier is $1.50 per million input tokens and $7.50 per million output tokens, flat across the 1 million-token context window. On the free tier, a line reads Used to improve our products: Yes.
AI Studio required a terms acknowledgment. The direct link into Google AI Studio with Hydrogen preselected lands on a first-use page that asks for an 18-and-over developer acknowledgment. We don't accept terms of service on our test accounts, so we stopped there and read the public docs instead. The model page lists the context window (1,048,576 tokens) and the input types.
We didn't test a free consumer account in the Gemini app, so we can't confirm what the picker looks like on a free plan. Google's announcement footnote says free users get Hydrogen via auto model selection when it's the appropriate model for the prompt, so we can't confirm that free-plan users can select Hydrogen manually.
The honest catch
The API free tier allows Google to use prompts and responses to improve its products. If that's a concern, the paid tier is the alternative, or another model entirely.
The knowledge cutoff is March 2026, per the model card, so recent events may be missing. The model card lists hallucinations, occasional slowness, and timeout issues as limitations, and says the model may use more tokens at higher thinking levels.
The benchmark numbers come from Google's announcement. Google picked the benchmarks, and we didn't run an independent test.
And for the hardest problems, Google positions Carbon and Argon for deep, long-horizon reasoning and Hydrogen for rapid execution across everyday coding and agentic workflows.
The full walkthrough
Our study guide covers the requirements first (what's free, what needs an account, what each plan gets), then a model-picker walkthrough you can run on your own account, the API and AI Studio routes, and a dated record of what we did and didn't test: Gemini 4 Hydrogen: Google's everyday model, free to try.
FAQ
What is Gemini 4 Hydrogen?
It's a Google model released on November 12, 2026, the third member of the Gemini 4 family alongside Carbon and Argon. Google positions it for rapid execution on everyday coding and agentic workflows, while Carbon and Argon handle deep, long-horizon reasoning on harder problems.
Is Gemini 4 Hydrogen free?
On the Gemini API, yes: the pricing page lists input and output as free of charge on the free tier, with the paid tier at $1.50 per million input tokens and $7.50 per million output tokens. In the Gemini app, Google's footnote says free users get Hydrogen via automatic model selection when it's the right model for the prompt. Our paid work account showed Hydrogen in the model picker; we didn't test every paid plan.
How is Hydrogen different from Gemini 4 Carbon?
Hydrogen is adapted from Carbon but tuned for speed on everyday work. Google's Carbon announcement positions Carbon for complex, multi-step coding, debugging, and scientific research, and says it uses more tokens than Argon to explore, reason, and verify. Google's pricing page on November 10 listed Carbon as excluded from the API's free tier. Hydrogen is the everyday model with a free tier.
What can I use Gemini 4 Hydrogen for?
Google highlights writing, research, and everyday work (drafts, study guides, rewrites, documents, clear tables), everyday coding, and agentic workflows where the model calls tools and follows instructions across multiple turns. Google says Hydrogen is also available in Search, where it's intended to handle multi-part and typo-filled queries better.
Does the free API tier use my prompts to improve Google's products?
The pricing page lists Hydrogen input and output as free of charge on the API's free tier, though we didn't run an API request ourselves. In the Gemini app, Google says free users receive Hydrogen through automatic model selection when it's appropriate for the prompt. The free tier uses your data to improve Google's products; the paid tier doesn't.
More on this: Gemini 4 Carbon Is Here: What It Is and Who Gets It and What Is a Token in AI? The Unit Behind Every Price and Limit.