Google opens its newest AI, Gemini 4 Carbon, to the public at a low price

Google has released Gemini 4 Carbon, which it calls its strongest AI model so far. It first went to a small group of cybersecurity defenders two weeks ago. Now paying developers and businesses can use it, and Google is charging far less than its main rivals. That low price is temporary.

What changed

Google says Carbon can compete with the best systems from Anthropic and OpenAI. It is built for hard work, such as complex coding with many steps and open-ended scientific research. Google says its other top model, Argon, launched only last month and will likely stay the better pick for most software and office-style knowledge work.

Google is also previewing a "swarm" feature in Antigravity, its platform for building AI agents. Agents are AI programs that carry out tasks on their own. Swarm is meant to turn one prompt into a coordinated team of agents.

What it costs

AI tools bill by tokens, which are small chunks of text the model reads or writes. Carbon launches at $2 per million input tokens and $10 per million output tokens. Google's own numbers put that at about a fifth of OpenAI's GPT-6 Astra and a quarter of Anthropic's Fable 5.2.

The introductory period ends at the end of November, and prices then double. Carbon would still be well below those rivals per token. But models can use different numbers of tokens for the same job, so a lower price per token does not always mean a lower bill.

Who can use it

• Paying developers and businesses, through the Gemini API in Google AI Studio and Gemini for business.
• Google AI Ultra subscribers, in the Gemini app, AI Mode in Search and Antigravity.
• Lower tier and free users: it is not clear when, or if, they will get it.

What to keep in mind

Google's benchmark results show Carbon near the top on most measures, but behind Fable 5.2. Benchmarks are standard tests, and companies can pick the results that flatter their tools. A business should try Carbon on its own real tasks before drawing conclusions.

The release follows Argon by only a few weeks. It marks a return for Google to the front of AI research, after months in which the lead looked like a two-way race between Anthropic and OpenAI.