LATIDIA · Ciberseguridad

CyberArena: Evaluación comparativa de agentes de IA en combate simétrico de ataque y defensa en vivo

arXiv: 2610.19005v1Tipo de anuncio: nuevo Resumen: los agentes de IA están demostrando capacidades de ciberseguridad ofensivas y defensivas cada vez más fuertes en el mundo real. Los puntos de referencia actuales evalúan principalmente a los agentes de forma aislada

La noticia

arXiv:2610.19005v1 Announce Type: new Abstract: AI agents are demonstrating increasingly strong offensive and defensive cybersecurity capabilities in the real world. Current benchmarks mostly evaluate agents in isolation against static targets. Real cyber operations, however, are dynamic and coupled: security postures constantly change as attackers and defenders adapt to each other in real time. Static benchmarks quickly saturate and miss an agent's ability to attack and defend simultaneously under live pressure. To address this gap, we present CyberArena, a dynamic, symmetric adversarial benchmark built around multi-round attack-and-defense combat. In isolated sandboxes, two agents compete under identical rules, acting concurrently as both attackers and defenders over continuously changing services. This converts static