Google launches Gemini 4 Hydrogen, a faster and cheaper model for everyday work

ModelsHUDERA AI team · 2026-11-12

On 12 November 2026 Google announced Gemini 4 Hydrogen, a model that sits below Gemini 4 Carbon and Gemini 4 Argon and is aimed at high-volume, repetitive work and at running agents. Through the API it costs $1.50 per million input tokens and $7.50 per million output tokens, a flat price across its full 1M-token context window. It has four thinking levels, including a new top level called xhigh.

Google says the model finishes tasks 2.5 times faster than Gemini 3.8 Flash and scores 57% on Terminal-Bench 4.1, against 18.2% for the older model. These are the company's own figures and have not yet been confirmed independently. The rollout is global: the Gemini API, Google AI Studio, the Gemini app and Google Search, plus Google Sheets and Antigravity for paid subscribers.

What does this mean for your business?If your company uses AI in volume, for customer replies, document summaries or data extraction, this is the tier of model that decides your monthly bill. Before switching anything, test it on a sample of your real work, especially in Arabic, and compare quality and cost with what you use now. The speed claims compare it with Google's previous model, not with competitors.

Sources

Summary written by the HUDERA AI team from the sources listed. Check the original source before any decision.

All news Current model rankings