Google releases Gemini 4 Hydrogen, a fast everyday model for agentic work

Google released Gemini 4 Hydrogen on November 12, 2026, as its fast model for everyday coding and agentic work. Google says it finishes tasks 2.5 times faster than Gemini 3.8 Flash and costs $1.50 per million input tokens. It is rolling out globally in the Gemini API, AI Studio, the Gemini app and Google Search.

What is Gemini 4 Hydrogen?

Hydrogen is the fast, cheaper member of the Gemini 4 family. Google keeps Gemini 4 Carbon and Gemini 4 Argon as its frontier models for deep, long-horizon reasoning. Hydrogen, by contrast, is "built for rapid execution across diverse agentic workflows," Google DeepMind's Koray Kavukcuoglu writes in the launch post.

The pitch is a division of labor. Teams still use Carbon or Argon for system design and subtle bugs, Google argues. But they also need "a fast, disciplined execution model for writing code, running tests, and verifying changes." Hydrogen fills that slot on its own, or next to the frontier models.

Google pitches it as an upgrade path from Gemini 3.8 Flash, and calls it a major standalone upgrade for teams migrating from Flash. For consumers, it now powers answers in Google Search and is in the Gemini app.

Hydrogen has four thinking levels: low, medium, high and a new xhigh tier. High is the default. Google says the levels scale in a predictable way, so you only wait and pay for the reasoning depth a task needs.

How fast and how smart is Gemini 4 Hydrogen?

On Google's numbers, it is far ahead of Flash and, on one broad index, level with Argon. The speed and Terminal-Bench figures come from Google's own testing. The Intelligence Index scores come from Artificial Analysis, an independent benchmark firm.

Speed first. At its default high setting, Google says Gemini 4 Hydrogen finishes tasks 2.5 times faster than 3.8 Flash. Early partners saw runs 2x to 2.8x faster with 12% to 23% fewer tokens, Google adds, on graph, SQL and data-heavy enterprise work.

The coding gain is larger. On Terminal-Bench 4.1, Hydrogen completes 57% of tasks, up from 18.2% for Flash. Google also says it cuts fabricated claims by more than 5x.

That is not the top of the table, though. VentureBeat notes that Google's own comparison puts Claude Sonnet 5.5 at 63.9% on the same test. So Hydrogen beats its predecessor by a wide margin, but not every rival in its class.

On the Artificial Analysis Intelligence Index, Gemini 4 Hydrogen scores 51 at its default setting, as Google reports in its post. That is one point above Gemini 4 Argon at 50. At xhigh, it reaches 54. VentureBeat reports Flash scored 31 at high reasoning.

Google lists more wins in its post. Hydrogen ranks fifth on Agent Arena, scores 58.8% on HLE-Diamond, and ranks first on Vals Tax Agent Bench at 52.3%. It also ties for first on the Vals Legal Research Bench at 55.8%. None of these have independent reruns yet.

How much does Gemini 4 Hydrogen cost?

Gemini 4 Hydrogen costs $1.50 per million input tokens and $7.50 per million output tokens. The price is flat across the whole 1 million-token context window, and the model can write up to 1 million output tokens.

The Gemini API pricing page adds the other tiers. Cached input costs $0.075 per million tokens. Batch and Flex cost half: $0.75 and $3.75. Priority costs $2.70 and $13.50. There is also a free tier.

This is why VentureBeat calls Hydrogen "pricier". Right now it costs twice Flash's introductory rate. But Flash moves to the same $1.50 and $7.50 on January 1, 2027. After that, Hydrogen costs the same as Flash per token, and Google's early partners report fewer tokens per task.

For earlier flagship prices, see our September model price roundup.

What is Orchestrator in Google Antigravity?

Orchestrator is a new Antigravity feature that switches between models, such as Carbon and Hydrogen, within one session. Instead of one model from start to finish, it hands each step to the model that fits it.

Google tested it on Terminal-Bench 4.0. Orchestrator reached near-Carbon task success, Google says, while using 24% less quota and finishing 19% faster on average. For parallel work, Carbon can act as the coordinator while Hydrogen runs worker and verifier agents in separate workspaces.

Individual developers need a paid Google AI Pro or Ultra plan to use Gemini 4 Hydrogen in Antigravity. Our Gemini 4 Carbon story covers the frontier model on the other side of that pairing.

G

Gemini

71

Google assistant wired into Workspace and Android

🇺🇸United States·Added 2mo ago

Freemium·from $4.99/mo·API

94trust score out of 100

G

Google Antigravity

85

Google's agent-first coding platform: a desktop agent manager, terminal CLI, Python SDK and IDE built around Gemini models.

🇺🇸United States·Added 29d ago

Freemium·from $30/seat/mo·API

84trust score out of 100

What changes in Google Search?

Google calls Gemini 4 Hydrogen "the biggest model upgrade to Search yet." Robby Stein, Google's VP of Product for Search, writes that it is now available globally in Search for all users.

There are three main changes. First, Search should handle long, multi-part questions, odd typos and very specific requests better. Second, Hydrogen improves query fan-out, the step that splits one question into many subqueries, so answers link to a wider range of sites. Third, responses get more images, structured answers, comparison tables and custom tools.

TechCrunch reports the same list and notes that the visual responses appear in AI Mode. Google says the improvements reach every language that AI Mode supports.

In the Gemini app, Google says Hydrogen follows instructions more closely and calls the right tools more often. Free users get it through auto model selection when it suits the prompt.

Where is Gemini 4 Hydrogen available?

Google says Hydrogen "is rolling out globally across our developer, enterprise, and consumer surfaces," starting November 12.

• Developers: the Gemini API in AI Studio and Android Studio, plus Google Antigravity for paid AI Pro and Ultra subscribers.
• Enterprises: Gemini for business, through the API or the model drop-down, plus Antigravity and Gemini in Google Sheets for eligible business plans.
• Everyone: the Gemini app model drop-down, Google Search for all users, and Gemini in Sheets, starting with AI Pro and Ultra subscribers.

On safety, Google reports Hydrogen cuts overreach in agentic tasks by up to 81% compared with Flash. It also says Hydrogen is up to 2x less likely to bypass permission boundaries on blocked tasks. The model card has the full list.

What this means if you build with AI

We checked Google's numbers against its launch post, its Search post and the Gemini API pricing page, and they match. The Flash index score and the Sonnet 5.5 row come from VentureBeat. Our take: Gemini 4 Hydrogen looks like the new default for high-volume agent work on Gemini, mainly because Flash's price rises to the same level in January. Still, Google's Terminal-Bench and Vals gains are its own claims, and Sonnet 5.5 scores higher on Terminal-Bench 4.1, per VentureBeat. If you are picking a coding assistant, run your own tasks first, and compare the products in our Claude and Gemini comparison. Our guide on how to choose an AI coding assistant covers the rest.

C

Claude

49

Assistant focused on long documents and code work

🇺🇸United States·Added 2mo ago

Freemium·from $17/mo·API

88trust score out of 100

C

ChatGPT

70

General assistant with tools, memory, files and voice

🇺🇸United States·Added 2mo ago

Freemium·from $8/mo·API

89trust score out of 100

2

What to watch next

Watch three things. First, independent reruns of Google's Terminal-Bench 4.1 and Vals results. Second, January 1, 2027, when Flash's introductory price ends and the cost gap closes. Third, whether Orchestrator moves beyond Antigravity into the Gemini API, which Google has not announced.