Gemini 4 Hydrogen is the fast, lower-cost model in Google’s Gemini 4 family. Google DeepMind released it on November 12, 2026, as the replacement for Gemini 3.8 Flash. Google positions it for everyday coding and agent work, alongside the larger Gemini 4 Carbon and Gemini 4 Argon, which it reserves for long and difficult reasoning.
Key specifications
Hydrogen accepts up to 1 million input tokens and, according to Google, can generate up to 1 million output tokens. Google says it works across PDFs, spreadsheets, charts, audio and recorded video as well as text and code. It offers four thinking levels: low, medium, high and xhigh. High is the default, and xhigh is new with this model. Higher levels trade speed and cost for more reasoning.
Pricing
The API price is $1.50 per million input tokens and $7.50 per million output tokens. The rate is flat across the whole context window. Ars Technica reports that this matches the price of 3.8 Flash. Carbon and Argon cost more per token, and their introductory price is scheduled to rise.
Availability
Google says Hydrogen is rolling out globally in Google AI Studio, the Gemini API, Android Studio, Gemini for business and Gemini in Google Sheets. Paid Google AI Pro and Ultra subscribers can use it in Google Antigravity. In the Gemini app it appears in the model menu, and Google says free users reach it through automatic model selection. It also powers AI responses in Google Search for all users.
Benchmarks Google reports
All figures below are Google’s own claims and have not yet been independently reproduced.
- Artificial Analysis Intelligence Index: 51 at the default high level and 54 at xhigh, against 50 for Argon.
- Terminal-Bench 4.1: 57 percent, against 18.2 percent for 3.8 Flash.
- HLE-Diamond: 58.8 percent.
- Vals Tax Agent Bench: 52.3 percent.
- Multimodal Knowledge Agent benchmark, introduced by Google the same day: 53.4 percent.
Google also says Hydrogen finishes tasks 2.5 times faster than 3.8 Flash at the default setting and makes more than five times fewer fabricated claims.
Strengths and limits
The main strength is price for capability: Google claims results near its larger models at the old Flash price. With the Orchestrator feature in Antigravity, Hydrogen can act as the execution model while Carbon plans. Google says this combination came close to Carbon’s task success on Terminal-Bench 4.0 while using 24 percent less quota.
The limits are those of the tier. Google still directs the hardest long-horizon problems to Carbon and Argon. The xhigh level produces more output tokens, so costs rise with the setting. Weights are not available, so the model cannot be run locally.
Alternatives
Within Google’s lineup, Argon and Carbon are the step up. For single-request classification and scoring, Google offers Gemini Decision 1.
Our coverage
See our report on the Gemini 4 Hydrogen launch, pricing and Search upgrade.