Announced 12 November 2026 · Unofficial briefing
Gemini 4 Hydrogen
Google’s fast execution model in the Gemini 4 family, built for everyday coding and agentic workflows. Frontier reasoning, in Google’s telling, stays with Carbon and Argon. This site only restates official published figures, each with a source and a date.
Paid API price
$1.50 / $7.50
Per 1M input / output tokens, flat across 1M context
Context window
1,048,576
Unified input, thinking, and output
Thinking levels
low · medium · high · xhigh
Default is high
Model ID
gemini-4-hydrogen
AI Studio, Gemini API, Agent Platform
What Google says it is
On 12 November 2026, Koray Kavukcuoglu announced Gemini 4 Hydrogen as the way Google is “bringing cutting-edge intelligence to everyone,” with speed and efficiency for everyday coding and agentic workflows. The same post contrasts it with Gemini 4 Carbon and Gemini 4 Argon, which “excel at deep, long-horizon reasoning on your hardest problems.”
The DeepMind model card adds that Hydrogen is adapted from Gemini 4 Carbon. It accepts text, images, audio, and video, with a combined input-output window of up to 1 million tokens. Default synchronous output is up to 64K tokens; up to 1M is available via background long decode. The knowledge cutoff is March 2026.
Source: Google Blog, 12 November 2026 · Source: Google DeepMind, 12 November 2026
Where it fits
Google’s developer documentation recommends Hydrogen as the upgrade path from Gemini 3.8 Flash and earlier Gemini 3 and 2.5 models, for everyday coding, interactive web apps, tool-using agents, multimodal document workflows, and parallel subagent execution. Carbon is described as the strongest frontier model; Argon as the model for most advanced long-horizon tasks with lower token use than Carbon. Hydrogen can also be paired with those two — Carbon planning, Hydrogen executing — including through a new Orchestrator option in Google Antigravity.
Source: Google AI Studio / Gemini API docs, Retrieved 13 November 2026
Source: Google Blog, 12 November 2026 · Source: Google Antigravity Blog, 12 November 2026
Read next
Access and use
API and AI Studio, the Gemini app, Search, Sheets, Antigravity, and Gemini for business — as Google described the rollout on 12 November 2026.
Pricing
Flat $1.50 / $7.50 per million tokens on the paid API, implicit cache, Cloud regional rates, and how that sits next to Carbon, Argon, and 3.8 Flash.
Comparisons
List prices from each vendor, plus Google’s published Hydrogen benchmark table versus Claude Sonnet 5.5 and GPT-6.1 Terra. No numbers invented here.
How people use it
Public experiments, third-party reviews, products, and videos. These are other people’s work, not tests by this site.
What this site will not do
- We do not claim affiliation with Google, and we do not use Google logos or official trade dress.
- We do not run formal benchmarks or qualitative scoring. If a published score appears, it is an official figure with a citation. On 14 November 2026 we also published original same-prompt tests (one run per prompt on the API free tier) with playable outputs and measured time, tokens, and equivalent paid cost — not scores. See Original tests.
- We do not invent user reviews, testimonials, or traffic statistics. Public third-party posts are collected on the home page and clearly labeled; we have not tested those items.
- We do not collect logins or payments.
Google Cloud currently describes Hydrogen as a short-term generally available release, with a minimum 45-day retirement notice. The full Safety & Security Report was listed as “coming soon” on the model card when we read it on 13 November 2026.
Source: Google Cloud documentation, Last updated 12 November 2026 · Source: Google DeepMind, 12 November 2026