Google launches Gemini 4 Hydrogen model, succeeding Flash 3.8
Google introduced Gemini 4 Hydrogen, a faster AI model that replaces the Flash 3.8 version and is being added to Search and its AI services.
Google has unveiled Gemini 4 Hydrogen, the latest model in its elemental AI series, which supersedes the Flash 3.8 version across the company’s products. Unlike the staged rollouts of the recent Argon and Carbon models, Hydrogen is available immediately to users, though access varies by subscription tier. AI Pro and Ultra subscribers can use the model in the Gemini app and AI Studio, while AI Plus and free users continue with the older Flash-Lite model.
According to Google, Hydrogen operates in four performance levels—low, medium, high and a new “xhigh” tier—and is up to 2.5 times faster than Flash 3.8 even in high mode. The company’s internal analysis says the model cuts “fabricated claims” by more than fivefold and improves code-generation scores, tripling Flash 3.8’s performance on the Terminal-Bench benchmark. In the Artificial Analysis Intelligence Index, Hydrogen achieved a score of 54 in xhigh mode, slightly above Argon’s 50.
Google positions Hydrogen as a tool for developers needing rapid, lower-cost code generation and testing, while Argon and Carbon remain aimed at longer-horizon tasks and higher security requirements. The model is also being integrated into Google Search, which the company describes as its biggest AI-search update to date, promising better handling of multipart queries and richer visual responses.
Pricing for Hydrogen’s API remains at $1.50 per million input tokens and $7.50 per million output tokens, unchanged from Flash 3.8. In contrast, Argon and Carbon are priced at $2/$10 per million tokens and are slated to increase to $4/$20 next month.