Google has launched Gemini 4 Hydrogen, a fast new artificial intelligence model that now powers its Search service for all users, in what the company describes as the biggest model upgrade to Search yet.
The model, announced on Thursday, is the new default in Google’s AI Mode and handles the hardest questions put to its AI Overviews. Robby Stein, Google’s vice-president of product for Search, said Hydrogen would give more helpful answers to complex questions, including “questions with multiple parts, hard-to-understand typos, or highly specific requests”.
Hydrogen is designed as the everyday workhorse of the Gemini 4 family, sitting below the deeper-reasoning Carbon and Argon models. Google says it is built for rapid execution across coding and so-called agentic workflows, where an AI system carries out multi-step tasks on a user’s behalf, and that it replaces Gemini 3.8 Flash as the company’s recommended fast model.
Google claims Hydrogen completes tasks 2.5 times faster than Gemini 3.8 Flash at its default setting and more than triples its score on the Terminal-Bench 4.1 coding benchmark, to 57 per cent from 18.2 per cent. On the Artificial Analysis Intelligence Index it scores 51 at its default setting, one point ahead of the larger Argon model, and 54 at its highest “xhigh” thinking level. Those figures are Google’s own, drawn from its own benchmarks, and have not been independently verified.
The model offers four thinking levels — low, medium, high and the new xhigh tier — so users and developers can trade reasoning depth against speed and cost. It has a 1 million-token context window, and Google is charging a flat rate of $1.50 per million input tokens and $7.50 per million output tokens through its API.
For Search, Google says Hydrogen improves its “query fan-out” technique, running several related searches at once behind a single question, which should surface a broader range of links, more visual answers and comparison tables. The upgrade applies across all languages supported by AI Mode.
Beyond Search, Hydrogen is rolling out from Thursday in the Gemini app, Google AI Studio, the Gemini API and Gemini for business, and in Google Antigravity for paid Google AI Pro and Ultra subscribers. Google is also introducing Orchestrator in Antigravity, which switches between models such as Carbon and Hydrogen within a single session; it says Orchestrator reached near-Carbon success rates on Terminal-Bench 4.0 while using 24 per cent less quota and finishing 19 per cent faster.
The launch lands in the middle of a fierce contest among the big AI labs to put their newest models in front of everyday users. OpenAI, Anthropic and Google have all shipped major releases in recent weeks, and the company whose model powers the world’s dominant search engine has the largest distribution of any of them. Whether Hydrogen’s promised gains show up in the quality of answers people actually get — and in the traffic publishers receive from Search — will only become clear once it has been running at full scale.
Sources: Google, “Introducing Gemini 4 Hydrogen”, https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-hydrogen/ ; The Verge, “Google’s new Gemini model is better at understanding complicated questions”, https://www.theverge.com/news/821356/google-gemini-4-hydrogen-ai-mode-search
