Gemini 4 Hydrogen
Google released Gemini 4 Hydrogen, a fast and cheaper Gemini 4 model for agentic and coding work, priced at $1.50 per million input tokens and $7.50 per million output tokens.
Hydrogen replaces Gemini 3.8 Flash and sits below the Carbon and Argon frontier models in the Gemini 4 family. Google says it finishes tasks 2.5 times faster than 3.8 Flash at its default high setting and more than triples the Terminal-Bench 4.1 score. It has a 1 million token context window with flat pricing and rolls out the same day across AI Studio, the Gemini API, the Gemini app, Search and Antigravity.
- Date
- Thursday 12 November 2026
- Lab
- Google DeepMind
- Kind
- model
- Access
- closed API
- Price
- $1.50 per million input tokens and $7.50 per million output tokens (OpenRouter batch listing $0.75 in / $3.75 out)
Figures
| Measure | Value | Measured by |
|---|---|---|
| Artificial Analysis Intelligence Index (default high thinking) | 51 Argon scores 50; xhigh setting reaches 54 | Artificial Analysis |
| Terminal-Bench 4.1 | 57% Gemini 3.8 Flash scored 18.2% | company |
| HLE-Diamond | 58.8% Google says it leads its class | company |
| Vals Tax Agent Bench | 52.3% Ranked #1 overall | Vals |
| Vals Legal Research Bench | 55.8% Tied for #1 | Vals |
| Agent Arena rank | #5 arena.ai live benchmark | arena.ai |
| Task speed vs Gemini 3.8 Flash | 2.5x faster at default high setting | company |
Google's blog describes it as replacing Gemini 3.8 Flash and dropping the Flash branding.
Sources
- blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-hydro
- deepmind.google/blog/introducing-gemini-4-hydrogen/
- techbuzz.ai/articles/google-launches-gemini-4-hydrogen-for-agentic-coding
- ghacks.net/2026/11/13/google-releases-gemini-4-hydrogen-in-search-and-the-gemini-app-repla
- mashable.com/tech/google-gemini-4-hydrogen-ai-model-launch
This record was checked and corrected against its sources on 13 November 2026. How we check