Google on November 12 made Gemini 4 Hydrogen available to developers and deployed it as the default model powering AI Mode in Google Search, extending the model's reach beyond coding into the company's core consumer product1,8. Robby Stein, VP of product at Google Search, said Hydrogen is now the new default model in AI Mode being used for the hardest questions in AI Overviews4. Koray Kavukcuoglu, SVP of Google DeepMind and Chief AI Architect at Google, said Hydrogen brings cutting-edge intelligence to everyone with speed and efficiency for everyday coding and agentic tasks7. Google dropped the Flash branding it had used for prior lightweight models, replacing it with the Hydrogen name6.
Benchmark, speed, and pricing
Hydrogen surpasses Gemini 4 Argon on the Artificial Analysis Intelligence Index at its default high thinking level, scoring 51 versus 50, and reaching 54 at its highest xhigh thinking setting. Google said the model completes tasks 2.5 times faster than Gemini 3.8 Flash, its previous fast model2. Google said Hydrogen often completes tasks faster than models in its class. Gemini 4 Hydrogen is priced flat across its entire 1 million-token context window at $1.50 per million input tokens and $7.50 per million output tokens.
The model handles repository coding, interactive web development, and multimodal workflows across four configurable thinking levels. It can run standalone or alongside frontier models Gemini 4 Carbon and Gemini 4 Argon. Google said Hydrogen delivers tighter instruction following and fewer fabricated claims to help developers ship production-ready code faster.
Search and consumer deployment
In Google Search, Hydrogen powers several improvements: better understanding of user intent, the ability to handle complex multi-part questions, more relevant links, and richer visual responses including structured tables and images. Google said these updates provide nuanced and accurate answers in all languages supported by its AI models. The Search upgrade is available globally.
Role alongside frontier models
Google framed Hydrogen as a complement to its frontier models: while Carbon and Argon handle deep architectural judgment such as system design or subtle bug resolution, Hydrogen serves as a fast execution model for writing code, running tests, and verifying changes.
Developers can access Hydrogen through the Gemini API in Google AI Studio and Android Studio, through Gemini for business, or via Google Antigravity for paid Google AI Pro or Google AI Ultra subscribers. Beyond developer channels, Google is also making Gemini 4 Hydrogen available in the Gemini app and Google Search. A feature called Orchestrator pairs Hydrogen with Gemini 4 Carbon inside Google Antigravity.
Business Insider reported on November 11 that Google had been letting staff test a fast new Gemini AI model named Hydrogen ahead of the public launch9.