aiqbench analysis

Anthropic is restructuring its external security initiatives, Project Glasswing and the Cyber Verification Program, into a unified three-tiered threat hunting framework. This consolidation aims to streamline how researchers and partners collaborate on identifying and mitigating advanced AI model vulnerabilities. The shift represents a significant evolution in Anthropic's approach to managing external intelligence for model safety rather than a new product launch or general corporate news.