Anthropic expands AI model access for government and cyber defense partners

Anthropic has introduced a new, enhanced version of its Cyber Verification Program (CVP), which provides advanced cyber capabilities and lowered blocking classifiers to qualifying security professionals. The program now offers three access tiers, each tailored to specific security needs.

The tiers are Defense Access for defensive work such as security operations, incident response, and vulnerability analysis, Red Team Access for authorized penetration testing and red-teaming, and Specialized Access for verified organizations authorized to test safety systems that impact people's lives or disrupt markets.

Organizations in the Defense Tier can use Claude Opus 5.5, Claude Sonnet 5.5, and Claude Mythos 5.1 models, while the Red Team Tier gains access to these models plus additional penetration testing capabilities. The Specialized Tier, currently available only to select organizations, has the fewest cyber blocks and is for testing safety systems of critical importance.

Qualifying organizations can apply through the program's website. Users can still employ generally available models for tasks like code review and vulnerability finding. Data retention is necessary for monitoring potential misuse, but zero data retention is an option for those using Claude Fable 5.1 or Claude Mythos 5.1 with the appropriate safeguards until a new solution, Enterprise Frontier Safeguards, becomes available later this fall.