Google introduced Gemini 4 Hydrogen on Nov. 12. It is a faster model in the Gemini 4 family, aimed at everyday coding and agentic work. Google says it is rolling out globally today.
What it is
Google positions Hydrogen next to its two frontier models. Carbon and Argon are for deep, long-horizon reasoning, Google says, while Hydrogen is "built for rapid execution across diverse agentic workflows." Our earlier stories cover Carbon in the API and for AI Ultra and Argon in the Gemini app.
Google also describes Hydrogen as a major upgrade for teams moving off Gemini 3.8 Flash.
Who gets it and where
Google lists these places:
- Developers: the Gemini API in Google AI Studio, and the paid tier in Android Studio.
- Enterprises: Gemini for business through the API or the model drop down, plus Gemini in Google Sheets for eligible business subscriptions.
- Google Antigravity: paid Google AI Pro and Ultra subscribers.
- Everyone else: the Gemini app model drop down and Google Search, for all users. Gemini in Google Sheets starts with AI Pro and Ultra subscribers.
Google says "all tiers" of AI Studio and the Gemini API are covered.
Price and thinking levels
Google prices Hydrogen flat across its 1 million-token context window: $1.50 per million input tokens and $7.50 per million output tokens. Google says it accepts up to 1 million tokens of input and can return up to 1 million output tokens.
It has four thinking levels: low, medium, high (the default) and a new xhigh tier. Google says you "only wait and pay for the reasoning depth a task requires."
Orchestrator in Antigravity
Google also introduced Orchestrator in Antigravity, which switches between models such as Carbon and Hydrogen within one session. Google says that on Terminal-Bench 4.0 it reached near-Carbon task success while using 24% less quota and finishing 19% faster on average. For parallel agents, Carbon can coordinate while Hydrogen runs worker and verifier agents.
Google's benchmark claims
These are Google's results, not ours:
- On the Artificial Analysis Intelligence Index, Google says Hydrogen scores 51 at its default high level against Argon's 50, and 54 at xhigh.
- Against Gemini 3.8 Flash, Google says Hydrogen finishes tasks 2.5 times faster at its default setting and scores 57% on Terminal-Bench 4.1 versus 18.2%.
- Google says it ranks #5 on arena.ai's Agent Arena and leads its class on HLE-Diamond at 58.8%.
Google also says it cuts fabricated claims by more than 5x versus 3.8 Flash. We have not seen how Google measured that.
Search and the Gemini app
Google says Hydrogen is the biggest model upgrade to Search yet, available globally to all users. Its Search post lists better handling of multi-part questions and typos, more visual answers, and an improved query fan-out that links to a broader range of sites. It also lists richer answers across supported AI Mode languages.
In the Gemini app, Google says Hydrogen follows tone and format more closely, calls tools more often and pulls context from Gmail, Google Drive and connected apps such as Airtable and Dropbox.
What isn't announced or confirmed
- Gemini app restrictions. The app line in Google's post carries a footnote we could not read, so tiers, regions and rate limits there are unconfirmed.
- Free-tier limits. Google does not state them in the posts we read.
- Whether older models, including Gemini 3.8 Flash, are being retired.
- Independent testing. Every benchmark above comes from Google.
- Our own testing. We have not used the model.
What this means for you
If you build with the API, the $1.50 and $7.50 rates are the number to compare against your current model. Check the thinking level you set, since Google says cost and latency scale with it.
In the app, look for Hydrogen in the model drop down. Search needs no action. For more on the family, see our testing report on Carbon and Argon. More in AI apps.
What we checked
We read Google's three Nov. 12 posts: the launch post, the Search post and the developer post. We did not test the model. We did not read the model card or the Gemini API guide.
