# Google and Anthropic now release their strongest cyber capabilities only to vetted defenders

Google gave Fairwind partners Gemini 4 Argon without cyber guardrails before its general release, and Anthropic's expanded program gives verified red teams Claude Mythos 5.1 with blocking classifiers reduced or removed.

- Published: 2026-10-14T06:06:44.194Z
- Canonical: https://polylog.news/ai/2026-10-14/google-and-anthropic-now-release-their-strongest-cyber-capab
- Publisher: Polylog (AI desk)
- Section: tech
- Sources: [Google DeepMind](https://deepmind.google/blog/developing-gemini-4-argon-responsibly-and-safely/), [Anthropic News](https://www.anthropic.com/news/cyber-verification-program)

The two labs most focused on agentic coding now grant offensive-grade cyber capability only to vetted customers, not to every paying user. Google's safety post on Gemini 4 Argon describes a phased release. Trusted defenders in the Fairwind…

This story is for subscribers. Read it in full at https://polylog.news/ai/2026-10-14/google-and-anthropic-now-release-their-strongest-cyber-capab (subscription information: https://polylog.news/pricing).