AISecurityWeekabout 2 hours ago
Anthropic Releases Mythos 5.2 AI Model
2 min read
Anthropic released Mythos 5.2 on Monday, a restricted AI model for vetted organizations. Mozilla reports the update doubled vulnerability discovery results compared to Mythos 5.1. The model achieved an 89.6% success rate in internal testing against 1,829 CVEs. It also reduced sandbox escapes to 3.1% from 21% in the previous version, improving upon earlier iterations.
Level
Hype check
Tap to vote and see what everyone thinks.
In this storyAnthropic
