Google launches Gemini 4 Hydrogen with 2.5x faster task completion

Google has officially launched the Gemini 4 Hydrogen model, built for fast execution and efficiency in daily coding and agent workflows. It is priced at $1.50 per million input tokens and $7.50 per million output tokens, with a 1 million token context window and no tiered pricing.

In benchmarks, Hydrogen scores 51 on the Artificial Analysis intelligence index at its default high thinking mode, and up to 54 at the top setting. Compared with the previous Gemini 3.8 Flash, it completes tasks 2.5x faster and cuts made-up errors by more than 5x.

Google also introduced the Orchestrator scheduling feature on its Google Antigravity platform, which switches between models within one session and cuts quota use by 24%. The model is rolling out worldwide, available to developers and everyday users through the Gemini API, the Gemini app and more, with Google Search also getting the upgrade.