Google releases Gemini 4 Hydrogen for fast coding and search
The new model replaces Gemini 3.8 Flash and offers four thinking levels to improve speed and intelligence in agentic workflows and Google Search.

Google has released Gemini 4 Hydrogen, a model designed for speed and efficiency in everyday coding and agentic workflows. It replaces Gemini 3.8 Flash across the company's products. Unlike the staged, security-conscious releases of Carbon and Argon, Hydrogen is available immediately with relaxed access limits.
The model operates at four thinking levels: low, medium, high, and a new xhigh tier. At its high default setting, Hydrogen is 2.5 times faster than Gemini 3.8 Flash. Google claims the model reduces fabricated claims by more than 5x and more than triples the Terminal-Bench score of its predecessor.
Hydrogen is being integrated into Google Search, which the company describes as its biggest update to AI-powered search tools ever. The model is expected to better handle long, multipart questions and navigate difficult typos. Responses in AI Overviews and AI Mode will include more visual elements, and the model's fan-out search has been tuned to link to more sources.
Hydrogen is priced at $1.50 per million input tokens and $7.50 per million output tokens across its 1 million-token context window.
Key facts
- Gemini 4 Hydrogen replaces Gemini 3.8 Flash and is available immediately.
- Hydrogen is 2.5 times faster than 3.8 Flash at its high thinking level.
- The model costs $1.50 per million input tokens and $7.50 per million output tokens.
- Hydrogen's xhigh mode scored 54 on the Artificial Analysis Intelligence Index.
Three perspectives
Neutral
Google has released Gemini 4 Hydrogen, a model designed for fast execution in coding and search tasks, featuring four distinct thinking levels.
Positive
Hydrogen provides significant leaps in speed and intelligence, including a 5x reduction in fabricated claims and improved performance in Terminal-Bench.
Negative
While Hydrogen is faster, Google still positions Carbon and Argon as the superior models for deep reasoning and long-horizon tasks.
Your week in news, every Wednesday
The biggest stories of the week for your edition, in your inbox every Wednesday morning, your time. Free, and you can unsubscribe at any time.