India, April 13 -- Anthropic announced last week that its latest AI model, Claude Mythos Preview, has already found thousands of high-severity security vulnerabilities across every major operating system and web browser, including flaws that have gone undetected for decades. The company is not making the model publicly available. Instead, it has launched Project Glasswing, a restricted initiative that gives access to 11 launch partners and over 40 additional organizations working on critical software infrastructure.

The decision to hold back the model came directly from what Anthropic found during its own testing. "AI models have reached a level of coding capability beyond most skilled humans at finding and exploiting software vulnerabil...