Claude Mythos 5.2 Outpaces Competitors in Vulnerability Detection
Original source
Anthropic’s Claude Fable 5.2 blocks fewer security tasks, while Mythos gets better at finding bugs
BleepingComputer →Anthropic released Claude Mythos 5.2 to partners as its dedicated cybersecurity model, outperforming competitors in vulnerability detection. The model identifies 10.2% more vulnerabilities than Claude Opus 5.5 and 14.4% more than Mythos 5.1. A security firm found that roughly one in six bugs Mythos caught had eluded all other frontier models tested. Mythos handles specialized work like penetration testing and exploit generation, positioning it as the company’s most capable security tool.
Claude Fable 5.2, released to general users, implements heavier safeguards to prevent misuse. The model blocks penetration testing, exploit generation, and certain vulnerability research, automatically deferring to Opus 4.8 when detecting high-risk activity. Anthropic reduced false-positive blocks by approximately 60% versus Fable 5.1, addressing user complaints about overly cautious restrictions interfering with legitimate work. The improvements reflect efforts to balance safety and usability for general security professionals.
Read the full article
Continue reading at BleepingComputer →This is an AI-generated summary. Read the original for the full story.