Anthropic expands Cyber Verification Program with three new access tiers

Anthropic announced on Oct 6, 2026, the launch of an expanded version of its Cyber Verification Program (CVP). The new program provides qualifying security professionals with access to advanced cyber capabilities and reduced blocking classifiers. It now consists of three access tiers designed to match the scope of cyber work performed by security teams.

New Access Tiers

The updated program integrates previous initiatives, Project Glasswing and the original CVP, into a single offering. Each tier includes access to Anthropic's most capable models, including Claude Opus 5.5, Claude Sonnet 5.5, and Claude Mythos 5.1.

• Defense Access: This tier is for defensive work, such as security operations center tasks, incident response, reverse-engineering malware, and analyzing vulnerabilities. Qualifying organizations include security teams at companies, nonprofits, universities, and government bodies, as well as operators of critical infrastructure and open-source maintainers. Anthropic states it expects many organizations to qualify for this tier and aims to respond to applications within a few days.
• Red Team Access: This tier adds authorized penetration testing and red-teaming to the defensive uses. It is available to in-house red teams, government red teams, and security firms. Users in this tier can only perform adversarial testing against systems they are authorized to test. Real-time blocks remain on actions that could cause physical harm or mass disruption, such as deploying ransomware. Anthropic notes that applications for this tier are expected to take a few weeks to review and are currently limited to organizations, not individual researchers.
• Specialized Access: This tier has the fewest cyber blocks and is reserved for a limited set of verified organizations authorized to test safety systems that could impact people’s lives or disrupt markets, such as flight operating systems and power grids. Anthropic states that it currently reviews every organization in this tier in collaboration with the US government. Existing members of Project Glasswing will transition to this tier.

Data Retention and Safeguards

Data retention is required for organizations enrolled in the program to monitor for cyber misuse. Anthropic says that once Enterprise Frontier Safeguards (EFS) is available later this fall, eligible organizations will be able to store data in cloud infrastructure they control. EFS is described as a solution that combines the privacy of zero data retention with robust safeguards. Until EFS is available, organizations with access to Claude Fable 5.1 or Claude Mythos 5.1 with zero data retention can also use CVP with zero data retention.

Program Efficacy and Impact

Anthropic evaluated the efficacy of the CVP protections using CyScenarioBench, an evaluation measuring whether models can plan and execute multi-stage cyber operations. According to Anthropic, without CVP access, every task was blocked on the first prompt. In the Defense Access tier, 46 of 50 trials were blocked at some point, while four tasks succeeded. In the Red Team Access tier, no blocks occurred, and Claude Opus 5.5 successfully completed 34 of 50 tasks, which Anthropic says is effectively equivalent to the model’s 67.6% success rate with no safeguards applied.

Regarding the impact of previous programs, Anthropic reports that partners in Project Glasswing uncovered at least 129,000 verified software vulnerabilities between April and July 2026. Through its own open-source scanning efforts, Anthropic found an additional 5,500 verified software vulnerabilities between April and October 2026. Of the vulnerabilities found by partners, more than 33,000 have so far been rated as critical- or high-severity. Anthropic notes that this figure is likely an undercount based on survey data from a subset of partners.

Why it matters

The expansion of the Cyber Verification Program reflects Anthropic's effort to balance the dual-use nature of advanced AI models. By creating tiered access, the company aims to provide defenders with powerful tools while maintaining safeguards against malicious use. The integration of previous programs and the introduction of specialized access for critical infrastructure testing highlight a growing collaboration between AI providers and government entities to secure essential systems.

---

Sources

• anthropic.com: Expanding the Cyber Verification Program \ Anthropic

This article was drafted with AI assistance and checked against the sources above. Company claims are reported as claims. Cover image is AI-generated.