Google has launched its long-awaited Gemini 4 Argon AI model to a wider audience, The Verge reported on Tuesday. The model is rolling out to Pro and Ultra subscribers in the Gemini app and AI Mode in Search. Ultra subscribers also get it on Antigravity, Google's coding tool. The Verge said it is not clear when, or if, it will reach free users.

The release follows a limited launch on Sept. 30. At that time, Argon went only to a select group of Google's cybersecurity partners through its Fairwind Program, TechCrunch reported. The model was trained for defensive cyber work, and Google said it can "autonomously find, validate, and patch critical software vulnerabilities," TechCrunch reported.

Google claims Argon beats rival models from OpenAI and Anthropic on tests in coding, security, science, finance and law, The Verge reported. TechCrunch said Google cited the benchmarking startup Vals to show Argon leads its AI model index. Those are Google's own claims.

Google says it is already using Argon inside the company, The Verge reported. Uses include finding and patching security flaws and debugging a drone physics simulator for Wing, Alphabet's drone delivery company.

Google is pricing the model aggressively, The Verge said. It launches at an introductory price of $2 per million input tokens and $10 per million output tokens. Prices double at the end of November, when the introductory period ends.

Argon is Google's first major frontier release since Gemini 3 in late 2025, The Verge reported. Since then, models from Anthropic and OpenAI left Google behind. Internal shakeups also raised doubts about its plans. Argon landed a day after Meta announced its own frontier model.

What we know: Gemini 4 Argon is now available to paying Gemini subscribers and developers at a low introductory price.
What we don't: How Argon performs in independent tests, and whether free users will get it.