WBench: A Comprehensive Multi-turn Benchmark for Interactive Video World Model Evaluation Paper • 2605.25874 • Published May 25 • 106
AgentHazard: A Benchmark for Evaluating Harmful Behavior in Computer-Use Agents Paper • 2604.02947 • Published Apr 3 • 19
PixelSmile: Toward Fine-Grained Facial Expression Editing Paper • 2603.25728 • Published Mar 26 • 118