Mozilla adopts Anthropic's Mythos AI as Firefox bug fixes surge
Mozilla just leveled up its security by using Anthropic's Mythos AI in its Firefox security research.
Since Anthropic unveiled Mythos in April 2026, Mythos has found thousands of serious bugs, including some that had been hiding out for over 15 years.
Thanks to this, Firefox jumped from fixing just 31 bugs in April 2025 to an impressive 423 a year later.
Mythos caught complex sandbox vulnerabilities
Mythos didn't just spot old issues: it also caught complex "sandbox" vulnerabilities that even top human researchers missed (despite Mozilla's $20,000 bug bounties).
While the AI suggests fixes, Mozilla engineers still review every patch before rolling it out.
The company says tools like Mythos are giving defenders a real edge, though they're staying cautious about how this tech could be misused.