Google launched Gemini 4 Hydrogen on November 12, 2026, a faster and cheaper model aimed at everyday coding and AI agents. Developers pay $1.50 per million input tokens and $7.50 per million output tokens, flat across a 1 million-token context window. Google says it is rolling out globally across the Gemini API, AI Studio, the Gemini app and Google Search. Here are the price, where it is available, what Google claims about speed and quality, and what has not been tested independently.
Gemini 4 Hydrogen API price
According to Google’s launch post, Hydrogen costs $1.50 per million input tokens and $7.50 per million output tokens, with the same rate across the full 1 million-token context window. The page does not state cached-token pricing, rate limits or quotas, and it does not label the model as a preview. Hydrogen supports up to 1 million input tokens and, per Google’s multimodal section, up to 1 million output tokens. It has four thinking levels: low, medium, high (the default) and a new xhigh tier.
Where Gemini 4 Hydrogen is available
- Developers: Google AI Studio on all tiers and the Gemini API, and Android Studio.
- Antigravity: paid Google AI Pro and Ultra subscribers.
- Business: Gemini for business and Gemini Enterprise, through the API or the model drop-down, and Gemini in Google Sheets for eligible business subscriptions.
- Consumers: the Gemini app model drop-down, where free users get Hydrogen through automatic model selection when it suits the prompt, and Google Search for all users globally.
Google’s post does not mention Vertex AI, so enterprise teams that buy through it should check Google Cloud’s own documentation. Fone Arena reports the same availability list and says the rollout is global from November 13.
How fast is it?
Google says Hydrogen finishes tasks about 2.5 times faster than Gemini 3.8 Flash at its default high setting. Partners in early testing reported runs 2x to 2.8x faster with 12% to 23% lower token use than 3.8 Flash. Gemini in Sheets generates output more than 30% faster than earlier models there. In Antigravity’s Orchestrator, a mixed Carbon and Hydrogen workflow on Terminal-Bench 4.0 used about 24% less quota and ran about 19% faster, with task success close to Carbon.
Benchmarks Google reports
- Artificial Analysis Intelligence Index: 51 at high and 54 at xhigh, against 50 for Gemini 4 Argon.
- Terminal-Bench 4.1: 57%, versus 18.2% for Gemini 3.8 Flash.
- HLE-Diamond: 58.8%.
- Vals Tax Agent Bench: 52.3%, first overall. Vals Legal Research Bench: 55.8%, tied for first.
- Agent Arena: ranked fifth. Agents’ Last Exam v1.1: 71.6%. Multimodal Knowledge Agent: 53.4%.
Most of these numbers come from Google. Fone Arena notes the speed and token figures are measured only against Gemini 3.8 Flash, partner results come from specific workloads, and the Orchestrator test used Terminal-Bench 4.0 while the model’s own score uses 4.1, so the two cannot be compared directly.
Safety claims
Google reports that, compared with 3.8 Flash, Hydrogen shows up to 81% less agentic overreach and 79% less under-service, and a 96.7% internal honesty rate. These are Google’s internal measures and the sources we read do not include independent audits. One inconsistency in the launch post: the text compares Hydrogen with Gemini 3.8 Flash, while some image captions name Gemini 3.5 Flash.
Who should look at it?
Hydrogen is positioned for high-volume coding and agent work, while Carbon and Argon stay Google’s models for deep reasoning. Teams already on 3.8 Flash have a clear upgrade path, since Google calls Hydrogen a “major standalone upgrade” for them. Before switching, run your own tasks: price per task depends on how many tokens the model uses, including thinking, and the launch post does not say how thinking tokens are billed.
Sources
- Google: Introducing Gemini 4 Hydrogen for agentic workflows
- Fone Arena: Google rolls out Gemini 4 Hydrogen
- Digit: Gemini 4 Hydrogen price and availability
Image: Asoundd / CC BY-SA 4.0 via Wikimedia Commons.
Like our coverage? Add stepphase.com as a preferred source on Google to see more of our stories in Top Stories.
