Google is completing its Gemini 4 lineup with Gemini 4 Hydrogen, following the Argon and Carbon models unveiled in October. According to the company, this new model is meant to bring “cutting-edge intelligence to everyone,” notably through its integration into its search engine. In a blog post published on Thursday, November 12, 2026, the Mountain View firm mentions “the largest model update ever made to Search.”
“More Useful and Intuitive” Answers in Google Search
Unlike Argon and Carbon, Gemini 4 Hydrogen is not designed to solve complex problems requiring deep reasoning. According to the California-based company, it is primarily aimed at the general public and is meant to “get things done more easily.” It offers several levels of reasoning, latency, and cost, designated by the labels “low,” “medium,” “high,” and “xhigh.” This gain in efficiency would not come at the expense of performance: “At every reasoning level, Hydrogen matches or surpasses some competing models in its category,” the announcement states.
In its default configuration, the model executes tasks twice as fast as Gemini 3.8 Flash, the company’s previous fast and low-cost model. It is also said to be less prone to hallucination, meaning inventing information. “It stays on track and follows instructions more faithfully across multi-turn conversations,” the announcement promises.
Its deployment in Google Search, launched on Thursday, November 12, 2026, is meant to allow the search engine to offer “more useful and intuitive answers.” Google cites four improvements. The first concerns understanding user intent, even when questions are fragmented, contain typos, or are very specific. The second relates to query fan-out, a mechanism that breaks down a query into several sub-topics processed in parallel, and is said to offer “a wider range of useful links than before.” The third is about presentation: answers incorporate more images, comparison tables, and structured responses, while loading just as fast as before. Finally, answers gain nuance in all languages supported by AI Mode, the conversational search mode deployed in France in July.
A Model Also Available in Gemini and via the API
Beyond Google Search, Gemini 4 Hydrogen is also being deployed this Thursday in the Gemini app. For writing tasks, the model follows the tone and rules set by the user more scrupulously, extracts contextual information from connected apps better, and “makes more use of the right tools.” “When you’re comparing products or synthesizing dense content, it produces clear tables and visual summaries without omitting essential details,” the announcement adds.
In the app, paying subscribers can select Hydrogen from the model selection menu. Google specifies that it will be automatically favored when it is particularly suited to a free user’s query. Developers, on the other hand, can access it via the Gemini API and Google AI Studio. In the Antigravity development environment, it is reserved for Google AI Pro and Google AI Ultra subscribers.
The model is priced at $1.50 per million input tokens and $7.50 per million output tokens, a token being a unit of text processed by the model. For comparison, Argon is priced at $2 input and $10 output until November 30, 2026, before a rate that, according to the source, is expected to double in early December.
