Google has launched Gemini 4 Hydrogen, a new AI model designed to deliver faster performance across coding, research and everyday tasks. The model began rolling out globally on November 12 across Google Search, the Gemini app, developer tools and business services.
Gemini 4 Hydrogen focuses on speed and efficiency while retaining advanced reasoning capabilities. Unlike Gemini 4 Carbon and Argon, which target more demanding problems, Hydrogen is optimized for writing code, debugging software, analyzing documents and completing multi-step tasks.
Google says Hydrogen completes tasks 2.5 times faster than Gemini 3.8 Flash at its default high reasoning setting. It also scored 51 on the Artificial Analysis Intelligence Index, compared with Gemini 4 Argon’s score of 50, reaching 54 at its highest reasoning setting.
In the Gemini app, Hydrogen improves instruction following, document creation, research and analysis of uploaded files. It can also work with connected services such as Gmail and Google Drive.
Google’s Search is receiving improvements designed to handle complex questions, discover more relevant websites and present answers using images, tables and other visual elements.
Developers can access Hydrogen through Google AI Studio and the Gemini API. Pricing starts at $1.50 per million input tokens and $7.50 per million output tokens, with support for a one-million-token context window.
Orchestrator is also a new addition to Google Antigravity, which automatically switches between Hydrogen and more powerful models during software development tasks to balance speed and computing costs.
Hydrogen is available through Google Search for all users, while free Gemini app users receive access through automatic model selection. Some developer and productivity features require eligible paid subscriptions.
While there are improvements in accuracy, efficiency and security, Hydrogen can still produce incorrect information, and its performance gains also vary depending on the task and reasoning settings.
Leave a Reply