1 paper
Yilun Zhao, Lujing Xie, Haowei Zhang +16
We introduce MMVU, a comprehensive expert-level, multi-discipline benchmark for evaluating foundation models in video understanding. MMVU includes 3,000 expert-annotated questions…