Anthropic Expands Cyber Verification Program for Security Professionals

• •Anthropic's Cyber Verification Program now includes three access tiers for security professionals.
• •Models available include Claude Mythos 5.1, Opus 5.5, and Sonnet 5.5 with reduced safeguards.
• •The program aims to balance the dual-use nature of AI in cybersecurity.

On October 6, 2026, Anthropic announced an expansion of its Cyber Verification Program (CVP), aimed at providing security professionals with broader access to advanced AI models for cybersecurity tasks. The updated program introduces three access tiers, allowing vetted security teams to utilize Claude Mythos 5.1, Opus 5.5, and Sonnet 5.5 with reduced safeguards for defensive and offensive operations, including penetration testing. The initiative is designed to balance the dual-use nature of AI capabilities, which can be employed for both security and malicious purposes. Anthropic's models have conservative safeguards to prevent misuse, but the new tiers are intended to empower defenders with the tools necessary to secure their systems effectively. Interested professionals can apply for access through the Anthropic website.