SignalTech/AI
Anthropic opens its most powerful AI models to more security teams
Anthropic expands access to its most powerful AI models for security teams.
Context
What happened: Anthropic is expanding its Cyber Verification Program, combining its Glasswing and original CVP initiatives into a three-tier program giving vetted security teams reduced safeguards on its most powerful AI models.
Key numbers: Glasswing partners found at least 129,000 verified vulnerabilities from April to July.
Why it matters: The April unveiling of Claude Mythos Preview raised fears AI could hack software before it was secured. The new tiers, vetted with the US government, widen access to Claude Opus 5.5, Sonnet 5.5 and Mythos 5.1.
What changed
Anthropic is expanding its Cyber Verification Program, combining its Glasswing and original CVP initiatives into a three-tier program giving vetted security teams reduced safeguards on its most powerful AI models.
Why it matters
The April unveiling of Claude Mythos Preview raised fears AI could hack software before it was secured. The new tiers, vetted with the US government, widen access to Claude Opus 5.5, Sonnet 5.5 and Mythos 5.1.