Gemini 4 Hydrogen: Google's Fast New AI Model Now Powers Search
Google launched Gemini 4 Hydrogen on 12 November 2026, a fast model for coding and agents that replaces 3.8 Flash. Price, benchmarks and where to use it.
Key points
- Google released Gemini 4 Hydrogen on 12 November 2026 as the fast, everyday model in its Gemini 4 family, succeeding Gemini 3.8 Flash.
- Gemini 4 Hydrogen now powers Google Search for all users and is rolling out in the Gemini app, Google AI Studio, Android Studio, Google Cloud and Google Antigravity.
- On the paid Gemini API, Hydrogen costs $1.50 per million input tokens and $7.50 per million output tokens, with a context window of about 1 million tokens.
- Google says Hydrogen finishes tasks 2.5 times faster than Gemini 3.8 Flash at the default setting and makes more than five times fewer fabricated claims.
- Hydrogen is designed to work alongside Gemini 4 Carbon and Gemini 4 Argon, which handle deeper, longer reasoning tasks.
Google launched Gemini 4 Hydrogen on Thursday, 12 November 2026, a fast AI model built for coding, AI agents and everyday tasks. It replaces Gemini 3.8 Flash as Google’s quick, lower-cost model and already powers Google Search for all users worldwide. Google says it is 2.5 times faster than the model it replaces, and it is rolling out in the Gemini app and to developers through the Gemini API.
What is Gemini 4 Hydrogen?
Hydrogen is the third model in Google’s Gemini 4 family, released a few weeks after Gemini 4 Argon arrived on 30 September and days after Gemini 4 Carbon. Google has split the work between them. In a statement quoted by Digital Trends, the company said: “Carbon and Argon do the deep thinking. Hydrogen gets the everyday work done, fast.”
The model accepts text, images, audio and video. It has four thinking levels, low, medium, high and a new “xhigh” level, with high as the default. Google also says it is better at choosing the right tool for a task, including pulling context from Gmail and Drive.
Where can you use it?
- Google Search: Hydrogen now powers Search for all users globally. Google calls it its biggest model upgrade for Search so far, with better understanding of what people mean, wider sets of links and more visual answers in AI Mode.
- Gemini app: rolling out now and selectable from the model menu. Free users get it when automatic model selection decides it suits the prompt.
- Developers: available in Google AI Studio on all tiers, in Android Studio and on Google Cloud, where VentureBeat reports it is listed as generally available.
- Google Antigravity: Google’s agentic coding tool offers Hydrogen to paid Google AI Pro and Ultra subscribers, along with a new “Orchestrator” mode that pairs Hydrogen with Carbon or Argon.
- Business: available to Gemini Enterprise customers through the API or the model menu.
How much does it cost?
On the standard paid Gemini API tier, prices per million tokens are:
| Gemini 4 Hydrogen | Price per 1M tokens |
|---|---|
| Input | $1.50 |
| Cached input | $0.075 |
| Output (including reasoning) | $7.50 |
The same price applies across the full context window of about 1 million tokens. Batch and Flex processing cost half. VentureBeat notes that this is more than Gemini 3.8 Flash costs today, because Flash is on an introductory rate until 31 December 2026, which is why it called Hydrogen “pricier.” For comparison, Argon costs $2 and $10 per million input and output tokens on its introductory pricing. For consumer plans, see PFM Today’s AI Subscription Price Tracker.
What do the benchmarks say?
All of these figures come from Google or were reported from its launch materials:
- Speed: finishes tasks 2.5 times faster than Gemini 3.8 Flash at the default setting.
- Accuracy: more than five times fewer fabricated claims than 3.8 Flash.
- Artificial Analysis Intelligence Index: 51 at the high setting and 54 at xhigh, compared with 50 for Argon and 31 for 3.8 Flash.
- Terminal-Bench 4.1 (coding in a terminal): 57%, up from 18.2% for 3.8 Flash. VentureBeat lists Anthropic’s Claude Sonnet 5.5 at 63.9% on the same test.
- Safety: up to 81% less “agentic overreach” than 3.8 Flash, meaning fewer cases of an agent going beyond what it was asked to do.
Digital Trends reported that Hydrogen beats OpenAI’s GPT 6.1 Terra and Claude Sonnet 5.5 on knowledge and multimodal tests but trails slightly on science, maths and computer use. Cloud storage company Box told VentureBeat its internal accuracy score rose from 74% to 79% with Hydrogen.
Why it matters
Fast, cheap models do most of the work in AI products. They answer search queries, run inside apps and carry out the many small steps of an AI agent. By putting Hydrogen into Search on day one, Google is using it at a scale few models reach. The Orchestrator approach, in which a large model plans and a fast model carries out the steps, is also how many companies now build agents to keep costs down. Google says Orchestrator came close to Carbon’s success rate on a coding test while using about 24% less quota.
The launch comes in a busy month for AI models, after Anthropic’s Claude Fable 5.2 release and price cut earlier in November.
What happens next?
Google says Hydrogen is also coming to Gemini in Google Sheets for AI Pro, Ultra and eligible business subscribers. Explicit caching for developers is not yet available, and Gemini 3.8 Flash’s introductory price ends on 31 December 2026, when its rate rises to match Hydrogen’s.
Frequently asked questions
What is Gemini 4 Hydrogen?+
Gemini 4 Hydrogen is Google's fast, efficient AI model for coding, agentic workflows, document and data work and everyday questions. It was launched on 12 November 2026 and replaces Gemini 3.8 Flash as Google's quick everyday model.
How much does Gemini 4 Hydrogen cost?+
On the standard paid tier of the Gemini API it costs $1.50 per million input tokens, $0.075 per million cached input tokens and $7.50 per million output tokens, including reasoning tokens. Batch and Flex processing cost half the standard rate. In the Gemini app, free users can get it through automatic model selection.
Is Gemini 4 Hydrogen better than Gemini 4 Argon?+
They do different jobs. Google positions Argon and Carbon for deep, long-horizon reasoning and Hydrogen for fast execution. Google says Hydrogen scored 51 on the Artificial Analysis Intelligence Index at its high setting, just ahead of Argon's 50, while costing less per token.
Sources
- Google Blog – Introducing Gemini 4 Hydrogen for Agentic Workflows
- Google Antigravity Blog – Gemini 4 Hydrogen and Orchestrator
- VentureBeat – Google launches Gemini 4 Hydrogen, a faster, smarter, pricier Flash successor
- Digital Trends – Google's next-gen Gemini 4 Hydrogen AI model is here
- FoneArena – Google rolls out Gemini 4 Hydrogen for developers, Gemini app with faster coding and Search upgrades
- Sundar Pichai
- Gemini 4 Hydrogen
- Gemini
- AI models
- AI agents
PFM Today Newsroom · News Desk
The PFM Today Newsroom tracks developments across startups, AI, design, business and more. Stories are researched from primary sources and reputable outlets, written with AI assistance, and reviewed against our editorial policy before publishing. Read our editorial policy.