Google launches Gemini 4 Hydrogen, a fast cheap AI model for everyone
A cheaper, faster Google model works now for free in the Gemini app and Google Search.
Google has officially launched Gemini 4 Hydrogen, a faster and cheaper AI model aimed at everyday coding and AI agent work 1. Unlike Google’s more powerful Gemini 4 family members, Hydrogen is open to all users — including those using the free versions of the Gemini app and Google Search .
For developers building apps with AI, the pricing is $1.50 per million input tokens (the text you feed into the model) and $7.50 per million output tokens (the response it generates). That is less than half the cost of Google’s flagship Gemini 4 Argon model 1. Hydrogen can process up to one million tokens of information in a single request — enough to handle long documents, lengthy conversations, or big blocks of code at once .
What you could do with it
- A free user of the Gemini app or Google Search could run a faster AI model for everyday coding and agent tasks at no cost 1.
- An AI app developer could build applications using Hydrogen for less than half the price Google charges for its flagship Gemini 4 Argon model 1.
- A coder could process up to one million tokens of text, such as long documents or large code blocks, in a single request 1.
Speed and accuracy
Google says Hydrogen completes tasks two and a half times faster than the older Gemini 3.8 Flash model 1. On rankings by the independent testing site Artificial Analysis, Hydrogen scores one point higher than Argon when both run at their default settings . Vals AI also called it the cheapest and fastest model among its top six ranked options .
In terms of reliability, Google claims the model makes up information five times less often than Gemini 3.8 Flash did 1. This matters because AI models sometimes “halucinate,” meaning they confidently state things that are not true .
Despite the improvements, Hydrogen does not lead every comparison. Artificial Analysis places its highest-level model one rank lower than Anthropic’s Claude Sonnet 5.5, while OpenAI recently cut GPT-6 prices with cheaper models of its own 1.
Built for everyday coding
Hydrogen is designed to handle routine, day-to-day tasks while Google’s bigger, more expensive models take care of heavier planning work 1. To make this easier, Google also introduced a new feature called Orchestrator inside its Antigravity coding tool. Orchestrator automatically switches between Hydrogen and the higher-end Gemini 4 Carbon model depending on what the task needs . Google says this setup delivers results close to using Carbon alone, while consuming 24 percent less of a developer’s usage quota .
In Google Search, Hydrogen is meant to tackle messy, multi-part questions more effectively and link readers to a wider variety of websites 1.
This story reports on a trend for information only. It is not advice or encouragement to do any of it.
How good is it
According to Google, Hydrogen completes tasks two and a half times faster than the older Gemini 3.8 Flash model 1. On rankings by the independent testing site Artificial Analysis, Hydrogen scores one point higher than Argon when both run at their default settings . Vals AI called it the cheapest and fastest model among its top six ranked options . However, Artificial Analysis places Hydrogen one rank lower than Anthropic's Claude Sonnet 5.5, and Google claims the model makes up information five times less often than Gemini 3.8 Flash did .
Corrections and updates
- Corrected on 12 November 2026. An earlier version said the model was from Google DeepMind specifically; the source refers to it as a Google product. An earlier version also claimed OpenAI lowered the cost of budget models alongside GPT-6; the source only states that OpenAI cut GPT-6 prices with cheaper models of its own, without separately mentioning budget models.
Key facts
- Costs $1.50 per million input tokens; free in the Gemini app and Google Search
- 2.5 times faster than Gemini 3.8 Flash on Google's measurements
- Can handle one million tokens of input at once
- Makes up facts five times less often than Gemini 3.8 Flash
- Trails Claude Sonnet 5.5 by one point on Artificial Analysis rankings
Sources
- Google's Gemini 4 Hydrogen brings a fast, cheaper model to everyone — thenextweb.com
If this story concerns you or your company, or uses something that belongs to you, ask for a correction or for it to be taken down. How requests work
Keep reading
How it worksOpenAI Releases 372 AI-Generated Math Proofs on GitHubHerald of AI
OpenAI Releases 372 AI-Generated Math Proofs on GitHub
The company is bypassing traditional journals, betting that formal verification can handle the flood of AI-produced mathematics.
OpenAI’s ChatGPT Wallet lets AI agents shop without seeing your card number
It marks the start of a new way to buy things online — your AI helper picks and pays, but you still approve every transaction and your real card stays hidden.
Business and rulesMirendil raises $1 billion to use AI agents for faster model developmentHerald of AI
Mirendil raises $1 billion to use AI agents for faster model development
A new startup founded by former Google DeepMind and Anthropic researchers says its automated platform lets small teams build large AI models quickly and cheaply.