AI Code Security with Claude Mythos Preview: Inside Tenable's 500+ Hours of Testing for Project Glasswing
Tenable, Thursday, August 6th, 2026
Tenable spent 500+ hours and 40 billion tokens testing Claude Mythos Preview for Project Glasswing.
Tenable invested more than 500 hours and 40 billion tokens testing Anthropic's Claude Mythos Preview as part of Project Glasswing.
The company's conclusion is that frontier AI will not run a code security program on its own.
Used well, however, it can make an existing program substantially stronger. Tenable reports that frontier AI dramatically scales security testing capacity relative to human-only review.