Anthropic says it found 5,500 verified vulnerabilities in April-October and Glasswing partners found 129K+ in April-July; 33K+ were critical or high severity (Reuters)

Anthropic reported finding 5,500 verified vulnerabilities between April and October. Concurrently, Glasswing partners identified over 129,000 vulnerabilities from April to July, with more than 33,000 classified as critical or high severity. This comes as Anthropic expands a program enabling vetted cybersecurity professionals to test its most powerful AI models with reduced safeguards, aiming to enhance security through extensive vulnerability discovery.

This report uniquely details the specific number of vulnerabilities Anthropic found (5,500) and the much larger scale of those found by Glasswing partners (129K+), unlike general reports on AI security programs.