Google Launches Gemini 4 Hydrogen: 2.5x Faster Than 3.8 Flash, Now in Search Worldwide

How it unfolded

Google announced in its official blog that Hydrogen is the Gemini 4 family's member for everyday work, complementing Carbon and Argon, which focus on deep, long-horizon reasoning. It reaches every entry point from developer tools to consumer products: the Gemini app model dropdown, the Gemini API, all tiers of Google AI Studio, Gemini for Business, and Antigravity for paid Pro and Ultra subscribers; the paid tier of Android Studio and Gemini in Sheets for eligible business subscriptions get it too.

Google Search is switching to Hydrogen at the same time, for all users globally. In the Search blog, Google calls it the biggest model upgrade to Search yet: the new model is meant to understand multi-part questions and questions with typos, and to return more visual, easy-to-scan answers and a wider range of links.

Performance and pricing

On the Artificial Analysis Intelligence Index, Hydrogen scores 51 at its default high thinking level, above the flagship Argon at 50, and reaches 54 at its highest xhigh setting. Against the previous 3.8 Flash, it finishes tasks 2.5 times faster at the default setting and scores 57% on Terminal-Bench 4.1, up from 18.2%, with Google saying fabricated claims fall by more than five times. All figures are Google's own test results.

Pricing is $1.50 per million input tokens and $7.50 per million output tokens, flat across the whole 1 million-token context window; input and output each go up to 1 million tokens. The model offers four thinking levels — low, medium, high (the default) and a new xhigh tier — with deeper reasoning costing more time and money as the level rises.

On other benchmarks, Hydrogen ranks fifth on Agent Arena, scores 58.8% on HLE-Diamond, ranks first on the Vals Tax Agent Bench at 52.3% and ties for first on the Vals Legal Research Bench at 55.8%, and scores 53.4% on the Multimodal Knowledge Agent benchmark.

Where it sits in the Gemini 4 family

After Carbon and Argon, Hydrogen is the third public model in the Gemini 4 family; Carbon opened to paying developers on Nov 10. Google positions Hydrogen as the fast model for everyday execution, to be paired with Carbon and Argon: Antigravity's new Orchestrator switches between Carbon and Hydrogen within a single session as needed, and Google says it reaches near-Carbon task success rates on Terminal-Bench 4.0 while using 24% less quota and finishing 19% faster.

The new model does not carry the Flash name. 9to5Google notes that Google dropped the Flash branding this time and named it Hydrogen outright; the model is positioned as the successor to 3.8 Flash for everyday, high-frequency tasks.

Key facts

1. Model: Gemini 4 Hydrogen, the third public model in the Gemini 4 family, built for everyday coding and agentic workflows
2. Speed: 2.5 times faster than 3.8 Flash at the default high setting; 57% on Terminal-Bench 4.1, against 18.2% for 3.8 Flash
3. Score: 51 on the Artificial Analysis Intelligence Index at the default setting and 54 at xhigh, against 50 for the flagship Argon at default
4. Price: $1.50 per million input tokens and $7.50 per million output tokens; 1 million-token context, with output up to 1 million tokens
5. Availability: From today across the Gemini app, API, AI Studio, Gemini for Business, Google Search (all users worldwide) and Antigravity (Pro and Ultra subscribers)
6. Thinking levels: Four: low, medium, high (default) and xhigh