GePBench: Evaluating Fundamental Geometric Perception for Multimodal Large Language Models

Shangyu Xing, Changhao Xiang, Xinyu Liu, Zhangtai Wu, Zhen Wu, Yue Yifan, Yuteng Han, Fei Zhao, Xinyu Dai
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:140106-140132, 2026.

Abstract

Geometric shapes play important roles in both physical world and human cognition. While multimodal large language models (MLLMs) have made significant advancements in visual understanding, their abilities to recognize geometric shapes and their spatial relationships, which we term geometric perception, are not explicitly and systematically explored. To address this gap, we introduce GePBench, a novel benchmark specifically designed to assess the geometric perception capabilities of MLLMs. Our extensive evaluations reveal that even the current state-of-the-art MLLMs exhibit significant deficiencies in geometric perception tasks. Furthermore, we show that models trained with GePBench data demonstrate considerable improvements on a wide range of downstream tasks, highlighting the critical role of geometric perception in enabling advanced multimodal applications. Our code and datasets are available at https://github.com/Changhao-Xiang/GePBench.

Cite this Paper


BibTeX
@InProceedings{pmlr-v306-xing26f, title = {{G}e{PB}ench: Evaluating Fundamental Geometric Perception for Multimodal Large Language Models}, author = {Xing, Shangyu and Xiang, Changhao and Liu, Xinyu and Wu, Zhangtai and Wu, Zhen and Yifan, Yue and Han, Yuteng and Zhao, Fei and Dai, Xinyu}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {140106--140132}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/xing26f/xing26f.pdf}, url = {https://proceedings.mlr.press/v306/xing26f.html}, abstract = {Geometric shapes play important roles in both physical world and human cognition. While multimodal large language models (MLLMs) have made significant advancements in visual understanding, their abilities to recognize geometric shapes and their spatial relationships, which we term geometric perception, are not explicitly and systematically explored. To address this gap, we introduce GePBench, a novel benchmark specifically designed to assess the geometric perception capabilities of MLLMs. Our extensive evaluations reveal that even the current state-of-the-art MLLMs exhibit significant deficiencies in geometric perception tasks. Furthermore, we show that models trained with GePBench data demonstrate considerable improvements on a wide range of downstream tasks, highlighting the critical role of geometric perception in enabling advanced multimodal applications. Our code and datasets are available at https://github.com/Changhao-Xiang/GePBench.} }
Endnote
%0 Conference Paper %T GePBench: Evaluating Fundamental Geometric Perception for Multimodal Large Language Models %A Shangyu Xing %A Changhao Xiang %A Xinyu Liu %A Zhangtai Wu %A Zhen Wu %A Yue Yifan %A Yuteng Han %A Fei Zhao %A Xinyu Dai %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-xing26f %I PMLR %P 140106--140132 %U https://proceedings.mlr.press/v306/xing26f.html %V 306 %X Geometric shapes play important roles in both physical world and human cognition. While multimodal large language models (MLLMs) have made significant advancements in visual understanding, their abilities to recognize geometric shapes and their spatial relationships, which we term geometric perception, are not explicitly and systematically explored. To address this gap, we introduce GePBench, a novel benchmark specifically designed to assess the geometric perception capabilities of MLLMs. Our extensive evaluations reveal that even the current state-of-the-art MLLMs exhibit significant deficiencies in geometric perception tasks. Furthermore, we show that models trained with GePBench data demonstrate considerable improvements on a wide range of downstream tasks, highlighting the critical role of geometric perception in enabling advanced multimodal applications. Our code and datasets are available at https://github.com/Changhao-Xiang/GePBench.
APA
Xing, S., Xiang, C., Liu, X., Wu, Z., Wu, Z., Yifan, Y., Han, Y., Zhao, F. & Dai, X.. (2026). GePBench: Evaluating Fundamental Geometric Perception for Multimodal Large Language Models. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:140106-140132 Available from https://proceedings.mlr.press/v306/xing26f.html.

Related Material