Google's Newest AI Model Is Now Open to Anyone Who Pays

Google has opened its newest model, Gemini 4 Argon, to everyone willing to pay for it. The company said the model became available globally on Oct. 13, 2026 in the Gemini app and AI Mode in Search for Google AI Ultra and Pro subscribers. It is also available on the paid tier of the Gemini API for developers in AI Studio and Android Studio, and for businesses through Gemini for business and Google Antigravity.

Google says Argon is built for long, multi-step work in fields including coding, finance, law, cybersecurity and science. The company states a limit of 1 million tokens for each request through its API, counting both what a user sends and what the model writes back. Tokens are the pieces of words a model processes.

Google's stated prices for reaching the model through its API are $2 per million input tokens and $10 per million output tokens. The company says that is an introductory rate, and that it rises to $4 and $20 after the introductory period expires on Nov. 30, 2026. Those prices cover the version of Argon intended for public access.

Google says that Argon scored 87% on the Embodied Reasoning benchmark, a test of spatial and physical problems, and that no other leading model scores higher. Every performance figure in Google's announcement is the company's own measurement and no independent evaluation of the model has been published.

The move ends a closed rollout. Google first gave Argon to cyber defenders on Oct. 1, 2026, and said at the time that it was not available to the public.

Two free routes stay open by application, both filled on a rolling basis. Google says it is offering 10,000 scientists the highest usage tier in Google Antigravity at no charge for six months, subject to eligibility, capacity and selection. The company says it is also expanding the Fairwind program, which carried the first release, to hundreds of additional partners.