Anthropic Grants Advanced AI Access for US Cyber Testing

Anthropic PBC is expanding access to its most advanced artificial intelligence models to a select group of verified organizations to test cyber capabilities. This initiative is conducted in collaboration with the United States government to evaluate the resilience of critical infrastructure.

The San Francisco-based company is providing access to Claude Opus 5.5, Claude Sonnet 5.5, and Claude Mythos 5.1. Authorized organizations will use these models to conduct high risk offensive testing on safety systems that protect essential services, including power grids, banks, and flight operating systems.