Google has released Gemini 4 Hydrogen, the third model in its Gemini 4 family. It is built for everyday work like coding, research, data analysis and writing, with a focus on speed and on finishing multi-step tasks without repeated instructions. Here is what changed, what it costs and whether it is worth switching to.
Table of Contents
What Is Gemini 4 Hydrogen?
Gemini 4 Hydrogen is Google’s fast, lower-cost model in the Gemini 4 lineup. Carbon and Argon handle deep, long-horizon reasoning, while Hydrogen is built for quick execution. It replaces Gemini 3.8 Flash in Google’s products, so it is the new everyday workhorse.
The name can be confusing because it says nothing about rank. Hydrogen is not a “lite” version of Argon. It plays a different role: speed and cost first, deep reasoning second.
When Did Google Release It?
Google announced Gemini 4 Hydrogen on November 12, 2026, just weeks after Argon and Carbon. Rumors about a “Hydrogen” model had circulated on X for days before Google confirmed it.
Key Features of Gemini 4 Hydrogen
Faster Responses and Four Thinking Levels
Google says Hydrogen finishes tasks 2.5 times faster than Gemini 3.8 Flash at its default high setting. It offers four thinking levels: low, medium, high (the default) and a new xhigh tier. That lets you trade speed for deeper reasoning depending on the job.
One-Million-Token Context Window
The model accepts up to one million tokens of input and can return up to one million tokens of output. Long documents, large code repositories and big data files can go in together instead of being split into pieces.
Better Coding and Agent Work
Coding is where Google puts the most emphasis. Hydrogen can find and debug bugs, run tests, search across several files at once and make focused edits instead of rewriting whole files. On Terminal-Bench 4.1, a test of command-line tasks, it scored 57 percent against 18.2 percent for Gemini 3.8 Flash. These are Google’s own numbers, so independent testing will give a fuller picture.
Research and Data Handling
Hydrogen also handles deep research and data analysis. It works with PDFs, spreadsheets, charts, audio and recorded video to produce reports and structured outputs.
Safety Improvements
Google says Hydrogen is more than 2.3 times more resilient to indirect prompt injection on the Gray Swan benchmark. It also reports a 96.7 percent score on its internal agent honesty evaluation, and says the model is up to two times less likely to bypass permission limits when a task is blocked. Prompt injection matters because AI agents often read websites and documents that may hide malicious instructions.
How Smart Is It Compared With Argon?
Despite its lower cost, Hydrogen scored 51 on the Artificial Analysis Intelligence Index at its default high setting, one point above Argon’s 50. At xhigh it reaches 54. Argon and Carbon are still positioned for deep, long-horizon reasoning, so benchmark scores don’t tell the whole story.
Gemini 4 Hydrogen Pricing
Through the Gemini API, Hydrogen costs $1.50 per million input tokens and $7.50 per million output tokens. The price stays flat across the full one-million-token window. Input tokens are what you send in, such as prompts and uploaded files. Output tokens are what the model writes back. Flat pricing helps if you regularly feed it very long documents.
Where You Can Use Gemini 4 Hydrogen
Google is rolling Hydrogen out across:
- The Gemini app
- Google Search
- Google AI Studio (see our complete guide to using Google AI Studio)
- The Gemini API and Android Studio
- Gemini for business
- Gemini in Google Sheets, starting with Google AI Pro and Ultra subscribers
- Google Antigravity, for paid Google AI Pro and Ultra subscribers
Access in the Gemini app depends on your plan. Google says free users get Hydrogen through auto model selection when it fits the prompt, rather than as a manual choice. According to Ars Technica, Hydrogen was available in the Gemini app and AI Studio for AI Pro and Ultra subscribers at launch, while free and AI Plus users were still on Gemini 3.5 Flash-Lite. Rollouts like this often widen over time, so check the model picker in your own Gemini app.
Gemini 4 Hydrogen in Google Search
This part affects the most people. Google made Hydrogen the default model in AI Mode for all users worldwide starting November 12. Google’s Search VP also said it handles the hardest questions in AI Overviews. If you publish content or track how often AI Mode cites your pages, note that citation data for November now spans a model change.
Orchestrator in Google Antigravity
Google also introduced Orchestrator in Antigravity, which switches between Carbon and Hydrogen within the same session. Carbon can plan and coordinate, while Hydrogen handles writing code, testing and verifying changes, and several Hydrogen agents can work on different parts of a project at once. On Terminal-Bench 4.0, Google says Orchestrator reached near-Carbon success rates while using 24 percent less quota and finishing 19 percent faster.
Gemini 4 Hydrogen vs Carbon vs Argon
| Model | Best for |
|---|---|
| Hydrogen | Fast, lower-cost everyday tasks, coding execution, agent workflows |
| Carbon | Complex reasoning, planning and coordinating larger projects |
| Argon | Deep, long-horizon reasoning on the hardest problems |
If speed and cost matter most, start with Hydrogen. If a task needs long planning or tricky reasoning, Carbon or Argon is the safer pick.
Should You Switch to Gemini 4 Hydrogen?
Developers running high-volume coding or agent workloads have the clearest reason to try it, given the speed gain and the price. Everyday users will mostly meet it automatically in Google Search. If you pay for Google AI Pro or Ultra, you can pick it manually in the Gemini app. For heavy reasoning work, test it against Carbon on your own tasks before moving over, since most benchmark figures come from Google itself.
Frequently Asked Questions
Is Gemini 4 Hydrogen free?
Partly. It is the default model in Google Search’s AI Mode for everyone, and it is available in Google AI Studio. In the Gemini app, free users get it through automatic model selection, while paid subscribers get it in the model picker. API use is billed per token.
What did Gemini 4 Hydrogen replace?
It replaces Gemini 3.8 Flash.
How big is its context window?
One million tokens of input, with up to one million tokens of output.
How much does the API cost?
$1.50 per million input tokens and $7.50 per million output tokens.
Is it better than Gemini 4 Argon?
It depends on the task. Hydrogen scored one point higher than Argon (51 vs. 50) on the Artificial Analysis Intelligence Index, according to Google. But Argon and Carbon are designed for deeper, longer reasoning, so test your own workload.
Sources: Google’s Hydrogen announcement, Google Search update, Search Engine Journal, Ars Technica
Read more: Complete Guide to Using Google AI Studio.




