NeurIPS 2025

MMPerspective: Do MLLMs Understand Perspective? A Comprehensive Benchmark for Perspective Perception, Reasoning, and Robustness

Figure: MMPerspective: Do MLLMs Understand Perspective? A Comprehensive Benchmark for Perspective Perception, Reasoning, and Robustness

Yolo Yunlong Tang, Pinxin Liu, Zhangyun Tan, Mingqian Feng, Rui Mao, Chao Huang, Jing Bi, Yunzhong Xiao, Susan Liang, Hang Hua, Ali Vosoughi, Luchuan Song, Zeliang Zhang, Chenliang Xu

First benchmark for perspective understanding in multimodal LLMs: 10 tasks, 2,711 images, 5,083 QA pairs across 43 models.

Paper Project page

First benchmark for perspective understanding in multimodal LLMs: 10 tasks, 2,711 images, 5,083 QA pairs across 43 models.

All publications