Back

Google opens its top AI model, Gemini 4 Argon, to its Pro and Ultra subscribers and to developers

Google's Gemini app screen with "Argon" chosen in the model menu of the Ask Gemini box, and Chat and Task buttons below it
Image: Google

Thirteen days after giving it first to security teams and early testers, Google has opened Gemini 4 Argon to subscribers of its paid AI Pro and Ultra plans and to developers who pay to build it into their own apps. An independent ranking of AI models on finance, legal, tax and coding work puts it first, just ahead of two Anthropic models. Google has not said whether or when people on its free plan or its cheaper AI Plus plan will get it.

Google on Oct. 13 opened Gemini 4 Argon, its new frontier model, to far more people. Frontier models are the most capable AI models, ones that match or beat the best that exist today. When Google announced Argon on Sept. 30, it rolled it out to a group of trusted cyber defenders and some early testers, and said everyone else would have to wait while it strengthened its safety limits. Now subscribers to its paid Google AI Pro and Ultra plans can use it in the Gemini app and in AI Mode, a feature of Google Search, and developers and businesses can use it through the paid version of Google's API, a connection that lets developers plug a service into their own apps.

Google AI Pro costs $19.99 a month and Ultra starts at $99.99 a month. In Antigravity, Google's app for building software with AI agents, AI that can carry out tasks on its own instead of only answering questions, Ultra is the only individual plan that gets Argon; businesses can use it there too. Developers pay per million tokens, the small chunks of text, often parts of words, that AI models read and write: an introductory $2 per million tokens Argon reads and $10 per million it writes. After Nov. 30, the price rises to $4 and $20. Through the API, Argon can handle up to about 1 million tokens in one job, what it reads and writes combined.

Google says Argon excels at long, complex, many-step work in coding, finance, law, cybersecurity and science. Vals AI, an independent company that tests AI models on finance, coding, legal and tax work, ranks Argon first on its overall index at 68.9 percent, just under two points ahead of Anthropic's Claude Sonnet 5.5 and Claude Opus 5.5. The lead is close: Vals shows each score can be off by about a point either way, and says no single model leads every part of its index.

Google calls Argon's safety and security protections its most comprehensive yet. It says outside testers who tried to get around its safety limits on biology could unlock less than 1 percent of the parts of the biology tests it runs, and that Argon is designed to refuse requests for help with cyberattacks. Trusted security teams get a version without those cyber limits through a program called Fairwind, which Google says it is now widening to hundreds more partners.

Google is also offering 10,000 scientists its highest Antigravity usage tier free for six months, to those who apply and are picked, subject to eligibility and capacity.

Google has not said whether or when people on its free plan or its $4.99 AI Plus plan will get Argon, or whether Antigravity subscribers below Ultra will. The introductory price lasts only until Nov. 30, so developers building on Argon now will pay twice as much per token after that. Google says more Gemini 4 models are coming, without giving dates.

More on Google