Anthropic opens its most powerful AI models to more security teams
The three-tier program gives vetted organizations access to Claude models with fewer safeguards after scans found more than 100,000 software vulnerabilities.
- On Tuesday, Anthropic expanded its Cyber Verification Program into three access tiers, opening Claude Opus 5.5, Claude Sonnet 5.5, and Claude Mythos 5.1 to a broader range of cybersecurity organizations.
- The revamped CVP integrates Project Glasswing, which provided Claude Mythos access to critical software teams, and the original Cyber Verification Program, combining two initiatives Anthropic operated for six months.
- Partners in the Project Glasswing initiative identified 129,000 verified software vulnerabilities between April and July 2026, with more than 33,000 rated as critical- or high-severity.
- In the Red Team Access tier, the model completed 34 of 50 tasks with no blocks, while the Defense Access tier saw 46 of 50 trials blocked, demonstrating graduated safety controls.
- Qualifying entities include security teams at universities, firms, and critical infrastructure operators, though the Specialized Access tier requires detailed vetting with U.S. government involvement to test safety-critical systems.
13 Articles
13 Articles
Anthropic opens its most powerful AI models to more security teams
Anthropic is expanding a program that allows vetted cybersecurity professionals to test its most powerful AI models with fewer safeguards, after its Project Glasswing initiative helped uncover more than 100,000 software vulnerabilities this year.
More than 33,000 vulnerabilities have so far been classified as critical or of high severity
Anthropic folds Project Glasswing into an expanded three-tier Cyber Verification Program
Anthropic PBC today expanded its Cyber Verification Program into three tiers of access for vetted security teams, with fewer blocks on cybersecurity work at each step up. The overhaul is aimed at a restriction Anthropic built into its own products. Because the company treats cybersecurity as dual-use, its generally available models carry conservative classifiers that […]
Crypto Briefing
Cryptocurrency News | Cryptocurrency Prices | Market CapAnthropic partners with US government to test advanced AI cyber capabilities
Anthropic's collaboration with the US government could boost its market position and valuation, influencing AI industry dynamics significantly.
The new programme has three levels, each with its own verification requirements and security controls.
Coverage Details
Bias Distribution
- 83% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
