Google is rolling out its new Gemini 4 Hydrogen model to Search globally, describing it as the biggest model improvement to Search so far. The upgrade is designed to better handle complex questions, including multi-part queries, highly specific requests and confusing typos.
While Google’s frontier models, Gemini 4 Carbon and Gemini 4 Argon, focus on deep reasoning for hard problems, Hydrogen is built for fast execution across agentic workflows. Google said it completes tasks 2.5 times faster than its previous fast model, Gemini 3.8 Flash, and outperforms Argon on the Artificial Analysis Intelligence Index at its default thinking level.
In Search, Hydrogen improves understanding of user intent, follows instructions more precisely and strengthens the fan-out technique that splits a prompt into multiple subqueries, helping users find relevant sites they may have missed. AI Mode will also show more images, structured responses, comparison tables and custom tools, with better answers across supported languages.
Beyond Search, Hydrogen can handle tasks such as writing code and verifying changes, or work alongside Carbon and Argon. It offers four thinking levels, including a new extra-high tier, and is available through AI Studio, the Gemini API, the Gemini app and Antigravity for paid subscribers.




