Anthropic Expanded Cyber Verification Program

The company has integrated its security initiatives to provide vetted professionals with access to specialized AI models.

Updated on Oct. 6, 2026 in Cybersecurity

Bold vector editorial illustration of a modular server rack with complex geometric cabling, representing cybersecurity infrastructure. — Anthropic has consolidated its Cyber Verification Program to provide vetted security teams with specialized access to advanced AI models for vulnerability testing. AI Illustration. Upload story photo >

Live Poll

Should tech companies provide security teams broader access to powerful AI models for software testing?

Anthropic has revamped its Cyber Verification Program by combining two existing efforts to help secure critical software globally. The initiative provides vetted security teams access to advanced AI models including Claude Opus 5.5, Sonnet 5.5, and Mythos 5.1.

Why it matters

By granting security professionals access to models with reduced safety safeguards, the company aims to accelerate the identification and remediation of software vulnerabilities. This collaborative approach allows for more efficient large-scale security testing.

The program features three distinct tiers: Defense, Red Team, and Specialized. Partners identified 129,000 verified vulnerabilities between April and July, with more than 33,000 categorized as critical or high severity.

The players

Anthropic

Anthropic is an AI safety and research company that develops large language models like Claude.

U.S. government

The U.S. government coordinates with private technology firms to vet security researchers and protect critical infrastructure.

The details

Anthropic now conducts member vetting in coordination with the U.S. government to ensure security controls are maintained during testing. The program represents a consolidation of various internal and external scanning efforts used throughout the year.

April 2026 marked the start of the vulnerability scanning period.
Verified vulnerability data collection concluded in July 2026.
Anthropic announced the revamped program on October 6, 2026.

The Tech Race

This program marks a shift toward specialized access for AI models, following the established trajectory of the Claude 3.5 Sonnet AI model. It reflects an effort to maintain a competitive security edge by allowing researchers to probe more advanced, less-constrained technology.

For security professionals and software developers, this program provides new avenues to utilize AI models for identifying bugs in their own systems. Increased identification of vulnerabilities could lead to more robust software updates and better protection against cyber threats for end users.

The takeaway

The move demonstrates a strategic shift toward democratizing high-level security tools for vetted experts. Developers should remain vigilant for increased software patching cycles as these specialized security teams identify more critical vulnerabilities.