Models
Your search engine gets Google's fast AI, which Google says performs like the big one
Gemini 4 Hydrogen, Google's fast model, now answers questions in Search for all users and in the Gemini app. Google says it performs almost as well as its most powerful models, but for now only Google is saying so.

The essentials
- Gemini 4 Hydrogen now works in Google Search for all users worldwide, according to the company (Google Blog, TechCrunch).
- Claimed by one partyAccording to Google, it finishes tasks 2.5 times faster than Gemini 3.8 Flash, the previous fast model (Google Blog, Ars Technica, TechCrunch).
- Claimed by one partyGoogle says it scores 51 points on an Artificial Analysis index that measures general ability, one more than Gemini 4 Argon (Google Blog, TechCrunch).
- Unlike Carbon and Argon, it is available from day one in the Gemini app, to businesses and to developers (Google Blog, Ars Technica, The Verge).
Why it matters
Until now, Google's newest models went first to specialists and to people paying for the most expensive subscription. With Hydrogen, the fourth generation of Gemini reaches anyone who uses Search, all at once. If Google's numbers hold up, the fast version would no longer be a second-tier AI and would perform close to the big ones. Someone outside Google still needs to check.
If you use AI Mode in Google Search, your answers now come from this model. You should see more tables, images and links, and it should also better understand long questions or ones with typos.
The details
Google unveiled Gemini 4 Hydrogen on November 12 (Google Blog). It is the fast model in the family. Its bigger siblings, Carbon and Argon, take their time with hard problems. Hydrogen is designed to handle one short task after another without delays (Google Blog, TechCrunch). According to Ars Technica, it replaces Gemini 3.8 Flash in the company's products (Ars Technica). Google also says it makes up false information more than five times less often than Flash (Google Blog).
In Search, Google promises that it understands questions with several parts, with typos or with very specific requests (Google Blog). It runs several searches at once on each part of the question to link to more websites. It also shows more images and comparison tables (Google Blog, TechCrunch). Robby Stein, head of product for Search, says Hydrogen is now the default model in AI Mode and is used in AI Overviews for the hardest questions (The Verge).
What we don't know
- No one outside Google has yet checked that it is that fast or that it makes fewer mistakes.
- Google does not say whether the model is also available in the Gemini app for free accounts.
Background
On November 11, we reported that Gemini 4 Carbon, Google's most powerful model, was opening up to developers and paying subscribers. According to Ars Technica, Carbon and Argon were rolled out in stages and at first only to cybersecurity researchers, while Hydrogen is available right away (Ars Technica).
Sources
- Introducing Gemini 4 Hydrogen
- Gemini 4 Hydrogen: everyday coding and agentic workflows for developers
- Google Search gets a major upgrade with Gemini 4 Hydrogen
- Introducing Gemini 4 Hydrogen
- Google rolls out Gemini 4 Hydrogen, finally replacing 3.8 Flash
- Google’s latest Search upgrade promises better answers to your most complicated questions
- Google’s new Gemini model is better at understanding complicated questions
- Google Releases Gemini 4 Hydrogen: 2.5x Faster Than Gemini 3.8 Flash
- Google cree haber dado con el santo grial de la IA: Gemini 4 Hydrogen es rápido, bueno y (casi) barato
- Gemini 4 Hydrogen 🚀, self-hosted Claude agents 🔒, Replit agent commerce 🛒
Was this story useful? Yes Not really