Gemini 4 Hydrogen
- Provider
- Modality
- MULTIMODAL
- Parameters
- —
- Context window
- 1,000,000 tokens
- Weights
- Proprietary
- Released
- 12 Nov 2026
Gemini 4 Hydrogen (12 November 2026) is the fast, low-cost model in Google's Gemini 4 family and the replacement for Gemini 3.8 Flash. It powers Google Search for all users and is in the Gemini app, the Gemini API and Google Antigravity.
Best for
- Everyday chat, search and quick answers
- Coding agents and tool calls where speed matters
- High-volume API work: summarising, extracting, classifying
- Long inputs, with a 1M-token context priced at one flat rate
How it compares — and the India angle
- Artificial Analysis scores it 51 at its default high setting, one point above Gemini 4 Argon and far ahead of 3.8 Flash's 31; at xhigh it reaches 54.
- Google says it finishes tasks 2.5 times faster than 3.8 Flash and reports 57% on Terminal-Bench 4.1 against 18.2%; that coding figure is Google's own.
- It costs twice 3.8 Flash's introductory price, but that price ends on 31 December, after which 3.8 Flash costs the same as Hydrogen.
- Free Gemini app users get it only through automatic model selection. Google's posts say nothing specific to India; the rollout is global.
Benchmarks
Representative public scores (approximate, higher is better) for relative comparison. Check the provider for the latest official results.
How to access
API gemini-4-hydrogen: $1.50 / 1M input and $7.50 / 1M output (thinking included), flat across the 1M context; cached input $0.075; batch and Flex $0.75 / $3.75. Free tier in Google AI Studio, where content is used to improve Google's products. No introductory discount.
Access Gemini 4 HydrogenFrequently asked questions
What is Gemini 4 Hydrogen?
Gemini 4 Hydrogen is a multimodal AI model developed by Google.
Who made Gemini 4 Hydrogen?
Gemini 4 Hydrogen was created by Google.
Is Gemini 4 Hydrogen open source?
No, Gemini 4 Hydrogen is a proprietary model with closed weights.
What is Gemini 4 Hydrogen's context window?
Gemini 4 Hydrogen supports a context window of 1,000,000 tokens.