Anthropic widens Claude access for security teams

Anthropic said it was widening its Cyber Verification Program to more security professionals. The programme gave vetted groups and individual researchers access to Claude’s strongest models with lighter safeguards for vulnerability research, malware analysis, incident response and penetration testing.

Anthropic said the public Claude models still blocked most of that work because the same capabilities could assist attackers. It said the programme was split into three access levels: Defense Access, Red Team Access and Specialized Access.

Defense Access was open to corporate security teams, government agencies, universities, critical infrastructure operators, open-source developers and individual researchers. Red Team Access was limited to organisations for approved attack simulations, and Specialized Access covered systems such as flight controls, power grids and banking infrastructure.

For the specialised tier, Anthropic said it vetted every applicant with the US government. It also said partners in the earlier Project Glasswing programme found at least 129,000 confirmed vulnerabilities from April to July 2026, with more than 33,000 classed as high-severity or critical.

Anthropic said those counts came from surveys of a subset of partners, and that the effect in practice was at least five times higher. It said several partners reported that the tool cut their work by months or years.