Mozilla says 271 vulnerabilities found by Mythos have "almost no false positives"

Mozilla said Anthropic’s Mythos AI found 271 Firefox vulnerabilities in two months, with almost no false positives.

Engineers credited the results to a better model and a custom harness that let the AI use Firefox testing tools like a human developer.

The system appears more practical than earlier AI bug-hunting efforts, which often produced unreliable reports.

The finding suggests AI-assisted security research could speed up vulnerability discovery when paired with specialized workflows.