GeoBench: Rethinking Multimodal Geometric Problem-Solving via Hierarchical Evaluation
Published in The Fourteenth International Conference on Learning Representations (ICLR 2026), 2025
We present GeoBench, a hierarchical benchmark for multimodal geometric problem solving with four reasoning levels and formally verified tasks.
Recommended citation: Yuan Feng, Yue Yang, Xiaohan He, Jiatong Zhao, Jianlong Chen, Zijun Chen, Daocheng Fu, Qi Liu, Renqiu Xia, Bo Zhang, Junchi Yan. (2026). "GeoBench: Rethinking Multimodal Geometric Problem-Solving via Hierarchical Evaluation." ICLR 2026. arXiv:2512.24119.
Download Paper