News
Expanding the Cyber Verification Program
Publication date not yet evaluated · Added · 4 reports anthropic.com
Page published
Coverage timeline
Why it matters
Anthropic's loosening of cyber safeguards for vetted teams signals growing use of frontier LLMs for offensive and defensive security work, a dual-use shift defenders should track as AI-assisted vulnerability discovery and red-teaming scale up.
Anthropic announced an expanded Cyber Verification Program (CVP) that gives vetted security professionals tiered access (Defense Access and Red Team Access) to its most capable Claude models (Opus 5.5, Sonnet 5.5, Mythos 5.1) with reduced cyber-blocking classifiers, merging the earlier CVP and Project Glasswing efforts. The post notes cyber capabilities are dual-use and retains real-time blocks on high-risk actions like deploying ransomware or damaging physical systems, with reporting tying Glasswing to large-scale automated vulnerability discovery.