Benchmark Radar
AI BENCHMARK PROFILE

StealthBench

General AIAgentsstealthbench.com

StealthBench measures operational stealth of autonomous offensive-security agents across six OPSEC dimensions, using 14 dockerized task scenarios and a three-model judge panel, with metrics like safe success rate and Stealth@Solve.

Released
2026-07-28
Readiness
Runnable
Primary field
General AI

Why it matters

It addresses the gap where agents find vulnerabilities but fail tradecraft, systematic across models. The benchmark supports development of stealth-aware agents and automated OPSEC monitoring for safe deployment.

Motivation

Stealth, the discipline of achieving an objective without revealing your presence, capabilities, or collected intelligence, is what separates sophisticated operators from detectable ones.

Primary resources

Benchmark Radar records only publicly supported details and links back to primary sources for verification.