Google Rolls Out Gemini 4 Argon to Vetted Cyber Defenders

· Original article ↗

Summary

Google is gradually giving vetted defenders access to Gemini 4 Argon, a model it says can find and patch vulnerabilities. Google also reported a flaw found in hospital software but did not identify the software or say if it was fixed.

Key points

  • Google is initially rolling out Argon to a select group of trusted cyber defenders through its Fairwind Program and to internal teams.
  • Google says Argon can autonomously find, validate, and patch critical software vulnerabilities.
  • Google says Argon found a critical vulnerability exposing sensitive personal information in hospital software; it did not identify the software or state whether the issue was addressed.
  • On CWE-bench v1, Argon tied for first with OpenAI’s GPT-6 Astra and xAI’s Grok 4.7, scoring 68%.
  • Google says Argon outperformed Gemini 3.8 Flash Cyber on its internal vulnerability benchmark and Wiz’s black-box penetration-testing benchmark.
  • Google says safeguards for broader access include refusing requests that could enable cyber or CBRN attacks, monitoring for misuse, and mitigations against indirect prompt injection and misaligned actions.

Article Details

Event Type
Google announced Gemini 4 Argon, a frontier AI model, and began rolling it out to a select group of trusted cyber defenders through its Fairwind Program.
Impact
Google says Argon can autonomously find, validate, and patch critical software vulnerabilities. In an early demonstration, it identified a critical vulnerability exposing sensitive personal information across healthcare software used by hospitals worldwide; the affected software and remediation status were not disclosed. Google says it is applying safeguards and a phased rollout before broader access.

Vendors

Products

Industries

Related Articles