Gemini 4 Argon has a 1M-token output limit, up from 64K for prior models; it initially costs $2/1M input and $10/1M output tokens, rising to $4 and $20 later (Matthias Bastian/The Decoder)

Google has launched Gemini 4 Argon, a new AI model designed to handle various tasks such as coding, research, and writing, with a focus on cybersecurity. According to Google, Argon can "autonomously find, validate, and patch critical software vulnerabilities." The model was trained specifically for defensive cyber work and has shown competitive scores across a range of tasks, including long-horizon software engineering and end-to-end business work.

Gemini 4 Argon has a 1M-token output limit, up from 64K for prior models, and initially costs $2/1M input and $10/1M output tokens, rising to $4 and $20 later, as reported by The Decoder. The model is currently being rolled out to a select group of Google's cyber partners through its Fairwind Program, with over 650 organizations already signed up, including CrowdStrike Holdings Inc. and Palo Alto Networks Inc.

Google is taking a phased approach to releasing Argon's capabilities, with the company still hardening safeguards against misuse for cyberattacks or weapons development. The model will eventually be available to Google AI Ultra subscribers, but for now, access is limited to vetted cybersecurity defenders and Google's own teams.

Brief written by urgent.news from Techmeme, TechCrunch, SiliconANGLE, Android Authority, XDA Developers — 5 reports on this story. Machine-written — may contain errors; check the original before relying on it.