The 2026 assessment, which began in May, simulated full intrusion chains—ranging from initial phishing entry to lateral movement and data theft. To mirror modern threats, researchers utilized AI-assisted development techniques to construct attack tools modeled on the MITRE ATT&CK Matrix. This year, the certification threshold was tightened to 92 percent, reflecting an industry-wide push to keep pace with evolving, automated offensive capabilities.
Eleven Security Suites Pass AV-Comparatives' Toughest AI-Driven Stress Test
With attackers increasingly deploying AI-built tooling to automate espionage, the latest AV-Comparatives Endpoint Prevention and Response test has set a new, higher bar for enterprise security. Out of 14 leading products evaluated against 50 complex, multi-stage attack scenarios, only 11 achieved the organization's highest Certified Leader status.

Bitdefender, Broadcom, Check Point, Elastic, ESET, Fortinet, G Data, Palo Alto Networks, TrendAI, VIPRE, and WithSecure secured the top ranking by meeting the rigorous combined threshold. While three vendors failed to reach the mark, AV-Comparatives withheld their names as part of its standard methodology. Andreas Clementi, CEO of AV-Comparatives, noted that while the current AI wave is new to the public, the defensive integration of machine learning has been a foundational element of endpoint security for two decades. The successful performers demonstrate that current enterprise tools remain capable of neutralizing threats built with AI-assisted code.



Comments (0)
No comments yet. Be the first!