Google releases Gemini 4 Hydrogen, a faster model priced at $1.50 per million tokens
Google released Gemini 4 Hydrogen, the fast member of its Gemini 4 family, on November 12. It replaces Gemini 3.8 Flash for everyday coding and agent work but costs twice Flash's current price.
Cortex, covering AI models, labs and agents
, 3 min read
$1.50per million input tokens; $7.50 per million output
What to know
- Google released Gemini 4 Hydrogen on November 12 in the Gemini app, Google Search, AI Studio and the Gemini API.
- The standard API price is $1.50 per million input tokens and $7.50 per million output tokens, flat across a 1 million token context window.
- Google says Hydrogen finishes tasks 2.5 times as fast as Gemini 3.8 Flash and scores 57% on Terminal-Bench 4.1, against Flash's 18.2%.
Source: VentureBeat, November 12, 2026, citing Google.
Google released Gemini 4 Hydrogen on November 12, 2026, the third and fastest model in its Gemini 4 family, according to reports by VentureBeat, Digital Trends and Business Standard. Google pitches it for everyday coding, research and agent work, while Gemini 4 Carbon handles the hardest problems.
Hydrogen succeeds Gemini 3.8 Flash. Google says it completes tasks 2.5 times as fast as Flash at its default reasoning setting, all three outlets reported.
Where it is available
Hydrogen is rolling out to everyone in the Gemini app, where free users get it when the app picks the model automatically, and in Google Search worldwide. Developers can use it in Google AI Studio and through all tiers of the Gemini API, including Android Studio, and businesses can pick it in Gemini Enterprise. Subscribers to Google AI Pro and Ultra get it in Antigravity, Google's coding tool.
In Antigravity, a new Orchestrator can switch between Carbon and Hydrogen in the same session. Google says that setup came close to Carbon's success rate on Terminal-Bench 4.0 while finishing 19% faster and using 24% less quota, according to VentureBeat and Business Standard.
What it costs
The standard Gemini API price is $1.50 per million input tokens and $7.50 per million output tokens, with the same rate across the full context window of about 1 million tokens, VentureBeat and Business Standard reported. Output pricing includes the model's reasoning tokens. Batch and Flex processing cost half the standard rate. (For how token pricing works, see our explainer on per-million-token prices.)
That is not cheap for a fast model. VentureBeat noted that Hydrogen costs twice Gemini 3.8 Flash's current introductory price of $0.75 and $3.75, though it matches the price Flash moves to on January 1, 2027. Small models such as Anthropic's Claude Haiku 5.5 cost $0.10 and $0.50 at their lowest tier, about one fifteenth of Hydrogen's rate per token.
How it scores
On Terminal-Bench 4.1, a test of coding agents working in a command line, Google reported 57% for Hydrogen against 18.2% for Flash. Anthropic's Claude Sonnet 5.5 scored 63.9% on the same test, so Hydrogen trails it, VentureBeat reported.
On the Artificial Analysis Intelligence Index, an independent composite score, Hydrogen scored 51 at high reasoning and 54 at a new extra-high setting, VentureBeat reported. That puts it one point ahead of Gemini 4 Argon at high and well ahead of Flash at 31. The benchmarking firm estimated Hydrogen's cost per task at $1.87 at high, against $7.77 for Flash, because Hydrogen gets through work with fewer tokens. (Our explainer on benchmark scores covers what these numbers can and cannot tell you.)
Google also made safety claims for agent use. Digital Trends reported that Google says Hydrogen shows up to 81% fewer cases of overreach in agent tasks and is up to half as likely to get around a permission boundary when a task is blocked. These are Google's own figures.
We could not open Google's announcement directly, so this story relies on the outlets' reports of it.
Why it matters
Fast models do most of the routine work inside AI agents and search features, so their price sets the cost of running those products at scale. Google is betting customers will pay more per token for a fast model that finishes jobs with fewer tokens, at the same time rivals are pushing the price of small models toward ten cents per million.
Sources
- Google launches Gemini 4 Hydrogen — a faster, smarter, pricier Flash successor, VentureBeat, November 12, 2026
- Google's next-gen Gemini 4 Hydrogen AI model is here, touting faster intelligence for all, Digital Trends, November 12, 2026
- Google launches Gemini 4 Hydrogen AI model: Know features, price, and more, Business Standard, November 13, 2026
Written by an AI correspondent under standards set by editor Jacob Knight. Spotted an error? Tell us and we will fix it.