Google launches Gemini 4 Hydrogen, a faster model replacing Gemini 3.8 Flash

Google has released Gemini 4 Hydrogen, a new AI model built for speed and efficiency that replaces Gemini 3.8 Flash across the company's products. The rollout covers developer, enterprise and consumer surfaces, and the model is also being integrated into Google Search. The launch comes as "gemini google" attracts strong search interest worldwide.

Hydrogen sits alongside Google's more powerful Gemini 4 Carbon and Gemini 4 Argon models. Google positions it as a fast execution model for everyday work, such as writing code, running tests and verifying changes, rather than a tool for the deepest reasoning tasks. Google's naming follows an elemental scheme, and the company describes the new model as a successor to the faster, more economical Flash models that many users rely on.

Speed, pricing and performance claims

Google says Hydrogen runs at four thinking levels: low, medium, high (the default) and a new xhigh tier. At the high setting, the company says the model completes tasks 2.5 times faster than Gemini 3.8 Flash. On Terminal-Bench 4.1, Google reports a score of 57%, compared with 18.2% for the previous Flash model.

On the Artificial Analysis Intelligence Index, Google reports that Hydrogen scores 51 at its default high setting, slightly ahead of Argon's 50, and 54 at xhigh. Google also says the model cuts fabricated claims by more than five times. These figures are company-reported and have not been independently verified in the material available to us.

Hydrogen is priced at a flat $1.50 per million input tokens and $7.50 per million output tokens across its full 1-million-token context window, according to Google's blog post.

Multi-model coding and Search changes

Google has also introduced Orchestrator in Google Antigravity, a feature that switches between models such as Carbon and Hydrogen within the same session. Google says Orchestrator reached near-Carbon task success rates on Terminal-Bench 4.0 while using, on average, 24% less quota and finishing 19% faster. In parallel-agent setups, Carbon can coordinate while Hydrogen handles worker and verifier tasks.

In Search, Google says Hydrogen improves understanding of long, multipart and highly specific questions, and can handle confusing typos better. Its fan-out technique, which splits a query into several subqueries, is intended to surface a wider range of websites. AI Overviews and AI Mode are expected to show more images, structured responses, comparison tables and custom tools. Google describes the update as the biggest improvement to its AI-powered Search tools to date.

Google has not provided independent confirmation of the Search gains. Some observers have noted that Google has previously made similar claims about source linking.

What happens next

Google says Hydrogen is available immediately across its developer, enterprise and consumer products, and that the Search update is rolling out to users globally. Earlier Gemini 4 releases were staged and initially restricted to higher-paying subscribers, but Google says limits for Hydrogen are far more relaxed. Hydrogen also ranks No. 5 on Agent Arena's benchmarks for multi-step tasks, according to the company.

For now, the key facts are clear: Hydrogen is live, it replaces Gemini 3.8 Flash, and its performance figures come from Google itself. Independent testing will show how the model performs in everyday use.