CyberNotableConfirmed
10.0
SentinelOne benchmark reveals frontier AI model limitations in malware investigation
SecurityWeek·3 days ago
Two AI platforms have reached human-level performance by achieving perfect scores at the International Mathematical Olympiad, marking a significant improvement over the 2024 silver-medal equivalent performance. This development demonstrates a rapid acceleration in AI reasoning capabilities, though the specific architectures and generalization limits remain proprietary or under analysis.