Standard

Gemini 4 Hydrogen Becomes Default Model in AI Mode, With More Visual Answers and Wider Links

By The Indus Pulse Editorial TeamUpdated Reviewed by an editor 3 min read
Illustration of an AI search interface with visual answers
AI illustrationCreated for The Indus Pulse with AI tools

Google said it has made Gemini 4 Hydrogen available to all Search users worldwide, with more visual and intuitive responses. The company called it the biggest model upgrade to Search yet. Robby Stein, VP of product at Google Search, said it is now the new default model in AI Mode and is being used for the hardest questions in AI Overviews, according to The Verge.

What Changes for Search Users

Google said Search can now handle more nuanced and complicated questions, quickly working out what the user needs. It also follows instructions more closely and calls tools more precisely, Google said, so that the response addresses each part of a question. Google said Hydrogen's reasoning improves its query fan-out technique, which runs simultaneous queries across the relevant subtopics to connect users to a broader range of links. The Verge reported that responses also include more visuals, such as images, comparison tables, graphs and maps. In one of Stein's examples, The Verge said, a user asks how phones work out which highway lane a car is in without cell service, and the AI responds with an interactive animated driving simulation. Google added that the model improves understanding of languages, with more nuanced and helpful responses across all supported AI Mode languages.

Availability, Pricing and Benchmarks

According to The Verge, the model is available in Search, the Gemini app, every tier of Google AI Studio, the Gemini API, Gemini for business and Google Antigravity for Google AI Pro and Ultra subscribers. Google said in its DeepMind blog that Gemini 4 Carbon and Gemini 4 Argon are its frontier models for deep reasoning on the hardest problems, while Hydrogen is meant for fast execution in agentic workflows. Google DeepMind lists flat pricing across the 1 million-token context window: $1.50 for every million input tokens and $7.50 for every million output tokens. Google said Hydrogen scores 51 to Argon's 50 on the Artificial Analysis Intelligence Index at the default high thinking level, reaching 54 at its xhigh setting. The Verge reported that Hydrogen offers thinking levels of low, medium, high and a new xhigh mode. Against Gemini 3.8 Flash, Google said, Hydrogen finishes tasks 2.5 times faster at its high default setting and more than triples Terminal-Bench 4.1 scores, 57% against 18.2%.

Orchestrator and Early Partner Testing

Google also introduced Orchestrator in Google Antigravity, which switches dynamically between models such as Carbon and Hydrogen during one session. Google said that on Terminal-Bench 4.0 Orchestrator came close to Carbon's task success rates while using 24% less quota on average and finishing 19% faster. In early testing, Google said, partners saw Hydrogen run 2 to 2.8 times faster while cutting token use by 12% to 23% against Gemini 3.8 Flash on complex graph, SQL and data-heavy enterprise workflows.

Spotted an error in this article? Tell us and we will check it.