In a groundbreaking evaluation conducted by AV-Comparatives, September 2026 witnessed the release of the seventh edition of its EPR (Endpoint Prevention and Response) test. This year, 11 out of the 14 participating security solutions achieved the prestigious Certified Leader status, the highest recognition from AV-Comparatives, thus underscoring their ability to counteract sophisticated cyber threats in an AI-driven landscape.
The EPR test critically assesses security products against a myriad of attack scenarios, reflecting the complex nature of modern cyber assaults. Launched in May 2026, the test included the examination of 14 top-tier enterprise security programs, all of which were challenged with 50 advanced attack scenarios designed with the help of AI-assisted tool development. This rigorous methodology ensures that the testing environment mirrors real-world conditions that organizations may face.
Among the tested security solutions, companies like Bitdefender, Broadcom, Check Point, and ESET emerged as frontrunners, each demonstrating exceptional protection capabilities and achieving a combined rating that surpassed the 92% threshold set by AV-Comparatives, which has risen from the previous standard of 90%. These results highlight the evolutionary strides made in the efficacy of cybersecurity technologies.
Earlier reports indicated concerns surrounding the increasing complexity and duration of cyberattacks, emphasizing an average attack lifecycle that spans approximately 241 days. Notably, the EPR test represents a proactive approach to such threats, employing modern strategies and leveraging findings from established frameworks like the MITRE ATTACK Matrix for Enterprises to simulate attack patterns as accurately as possible. This ensures that the security products tested are not only effective but also relevant to current threat landscapes.
As stated by Andreas Clementi, CEO of AV-Comparatives, the role of machine learning in cybersecurity has significantly evolved over the last two decades, allowing defenders to adapt and withstand the tactics deployed by cyber adversaries. For the first time, AI techniques informed the construction of attack scenarios in this year's test, combining robust existing methodologies with innovative advancements in threat analysis.
The certified solutions demonstrated solid responses against various attack vectors, including phishing attempts, privilege escalation, lateral movement, and data theft, reflecting realistic multi-step attack scenarios. Their performance indicates a well-developed defense mechanism capable of thwarting sophisticated AI-driven infiltrations that threaten organizational integrity.
While manufacturers of the three solutions that did not meet the certification criteria remain unnamed, this year’s bar of excellence serves as a wake-up call to all players in the cybersecurity field. By raising awareness of the evolving threat landscape and the importance of rigorous testing, AV-Comparatives aims to empower businesses to make informed choices about their cybersecurity investments.
For a thorough understanding of the report, including insights into CyberRisk Quadrants and detailed product analyses, the full findings can be accessed at
AV-Comparatives' official website.
AV-Comparatives is recognized as an independent benchmarking organization dedicated to the assessment of cybersecurity solutions, which have surpassed traditional antivirus software in capabilities. The organization is committed to delivering transparent and accessible insights that help organizations worldwide enhance their security postures. Their evaluations encompass a comprehensive range of security aspects, including endpoint protections, threat detection, and mobile security on diverse platforms such as PCs, Macs, Android, and IoT devices.