Gemini 4 Hydrogen: what changes in Search and AI Mode
Updated
Illustrative photo
A model built for messy, multi-part questions
Google says its new Gemini 4 Hydrogen model handles complex questions better. That includes requests with several parts, hard-to-read typos or very specific asks. The model splits a question apart and answers each piece in its response.
According to The Verge · Tech (US), Robby Stein, VP of product at Google Search, calls it the biggest model upgrade to Search so far. Stein says it is now the default model in AI Mode and handles the hardest questions in AI Overviews.
Answers get a wider range of links because the model runs simultaneous queries across related subtopics. Responses also carry more visuals: images, comparison tables, graphs and maps.
One example from Stein shows the AI building an interactive animated driving simulation. The question was how phones work out which highway lane a car is in where there's no cell service. Google also says language understanding improved across all supported AI Mode languages.
Where you can use it and the thinking levels
The model is available starting today in search, the Gemini app, all tiers of Google AI Studio, the Gemini API, Gemini for business, and Google Antigravity for AI Pro and Ultra subscribers. Users get low, medium, high and a new xhigh thinking level.
Google claims the model makes significant leaps in speed and intelligence over Gemini Flash 3.8. It also says it matches or beats some competing models in its class, such as Claude Sonnet 5.5 and GPT-6.1 Terra. Those are Google's own comparisons, not results I've checked.
Affiliate links: we may earn a commission from qualifying purchases, and it never changes our data or rankings. They open Amazon's search results: prices and availability are shown there.