Anthropic is expanding access to its newest Claude models for a broader range of cybersecurity teams by reducing restrictions. Through the company’s updated Cyber Verification Program, approved participants will gain access to Claude Opus 5.5, Claude Sonnet 5.5, and Claude Mythos 5.1. This follows the earlier launch of Project Glasswing, which was designed to secure critical global software. That initiative notably identified over 100,000 software vulnerabilities between April and July 2026, while Anthropic’s independent open-source scanning uncovered an additional 5,500 flaws from April to October.
Consequently, the artificial intelligence firm is rolling out a restructured Cyber Verification Program (CVP) that encompasses two distinct tracks. The first track, Glasswing, provides verified organizations with access to Claude Mythos, representing Anthropic’s most advanced cyber-capable model family.
The second track is the original CVP, where selected security groups are granted relaxed safeguards on Claude Opus and Sonnet models. Participants are categorized into three distinct tiers depending on their specific operations and access requirements.
These three tiers are designated as Defense, Red Team, and Specialized. The Defense tier supports tasks such as incident response and malware analysis, the Red Team tier is tailored for authorized security testing, and the Specialized tier features the highest level of access with minimal restrictions for select groups protecting critical infrastructure, which encompasses financial, aviation, and power networks.
Anthropic noted that participants undergo vetting prior to gaining access. The objective is to supply robust AI capabilities to qualified professionals for security applications while preventing potential misuse.
Also Read: Australia Weighs New AI Breach Rules as OpenAI, Anthropic Show Support
Fable vs Mythos: What is The Difference?
Fable 5.1 and Mythos 5.1 stem from the identical underlying architecture, with safety guardrails serving as the primary distinction. Fable incorporates stricter safety constraints and enjoys wider availability. Conversely, Mythos features fewer limitations tailored for vetted professionals operating within sectors like life sciences and cybersecurity.
This arrangement enables researchers to tackle actual security challenges without encountering premature model restrictions. For instance, security analysts often need to analyze dangerous code or evaluate system vulnerabilities, and Mythos is built to offer qualified users greater flexibility for those tasks.
While Anthropic is broadening availability to its more potent Claude models, access remains carefully tiered rather than universal. The strategy aims to empower trusted security personnel to detect software vulnerabilities more rapidly while maintaining strict oversight over these high-capability tools.
