Google Launches Gemini 4 Hydrogen Model to Replace Flash

Google is continuing its elemental artificial intelligence naming scheme with the rollout of the new Gemini 4 Hydrogen model. While these monikers are less descriptive than the Pro and Flash names of earlier releases, the purpose of Hydrogen is straightforward as it replaces the faster and more economical Flash models used throughout the Google ecosystem.

Unlike previous releases such as Argon and Carbon, which featured staged rollouts beginning with cybersecurity researchers or restricting access to high-tier subscribers, the limits for Hydrogen are much more relaxed with access beginning immediately. The new model has officially replaced Gemini 3.8 Flash across the company's products.

Performance and Pricing Improvements

According to Google, Hydrogen represents a major improvement over the previous Flash model. It operates in one of four thinking levels including low, medium, high, and the new xhigh mode. Even when running in high mode, it operates 2.5 times faster than Flash 3.8 while reducing fabricated claims by more than five times. Users will also see gains in code generation, with the model more than tripling 3.8 Flash's score in Terminal-Bench.

The model is live for AI Pro and Ultra subscribers, though AI Plus and free users remain on the older 3.5 Flash-Lite model. For developers connecting via the API, pricing is set at $1.50 per million input tokens and $7.50 per million output tokens, matching the previous rates of 3.8 Flash.