arXivAI Research
WorldAuditBench: Evaluating Multimodal Agent Capabilities in Interactive 3D World Auditing
WorldAuditBench：首個評估多模態 Agent 3D 虛擬環境審計能力的全新基準測試
WorldAuditBench is a new benchmark designed to evaluate how multimodal agents audit interactive 3D environments for physical and logical anomalies, revealing a massive gap between current AI and human performance.
2 min read