TY  - RPRT
TI  - OCR-Reasoning Benchmark: Unveiling the True Capabilities of MLLMs in Complex Text-Rich Image Reasoning
AU  - Mingxin Huang
AU  - Yongxin Shi
AU  - Dezhi Peng
AU  - Songxuan Lai
AU  - Zecheng Xie
AU  - Lianwen Jin
PY  - 2026
UR  - https://arxiv.org/abs/2505.17163
ID  - 2505.17163
ER  -