Google launches Gemini 4 Hydrogen for complex questions in Search and its apps

Google says Gemini 4 Hydrogen is designed to handle complex Search queries, including questions with several parts, typos or specific requirements. It is now the default model in AI Mode and is used for the hardest questions in AI Overviews, extending its reach across Google’s apps and services.

The model can split a query into subtopics and run simultaneous searches, which Google says helps it provide a wider range of links. Responses may also include visuals such as comparison tables, graphs and maps, or interactive simulations. It is available in Search, the Gemini app, AI Studio, the Gemini API and business services; users can choose low, medium, high or “xhigh” thinking levels. Google says it is faster and more capable than Gemini Flash 3.8 and competitive with some rival models.