MiMo-VL Technical Report

作者：X. Yue, Zhenrui Lin, Yi-Hao Song, Weikun Wang, Shu-Qin Ren, Shuhao Gu, Shicheng Li, Peidian Li, Liang Zhao, Lei Li, Kainan Bao, Hao Tian, Hailin Zhang, Gang Wang, Dawei Zhu, Cici, Chenhong He, Bowen Ye, Bowen Shen, Zihan Zhang, Zi-Ang Jiang, Zhixian Zheng, Zhichao Song, Zhen Luo, Yue Yu, Yudong Wang, Yu Tian, Yu Tu, Yihan Yan, Yi Huang, Xu Wang, Xin-dan Xu, X. Song, Xing Zhang, Xing Yong, Xin Zhang, Xia Deng, Wenyu Yang, Wenhan Ma, Weiwei Lv, Weiji Zhuang, Wei Liu, Sirui Deng, Shuo Liu, Shimao Chen, Shi-liang Yu, Shao-yang Liu, Shan-yong Wang, Rui Ma, Qiantong Wang, Peng Wang, Nuo Chen, Menghang Zhu, Kang Zhou, Kang Zhou, Kai Fang, Jun-Miao Shi, Jinhao Dong, Jiebao Xiao, Jiaming Xu, Huaqiu Liu, Hongsheng Xu, Hengxu Qu, Hao-Song Zhao, Hanglong Lv, Guoan Wang, Duo Zhang, Dong Zhang, Di Zhang, Chong Ma, Chang Liu, C. Cai, Bing Xia · 发表于：arXiv.org · 年份：2025 · DOI：10.48550/arxiv.2506.03569 · 被引用次数：52 · 研究领域：Computer Science

We open-source MiMo-VL-7B-SFT and MiMo-VL-7B-RL, two powerful vision-language models delivering state-of-the-art performance in both general visual understanding and multimodal reasoning. MiMo-VL-7B-RL outperforms Qwen2.5-VL-7B on 35 out of 40 evaluated tasks, and scores 59.4 on OlympiadBench, surpassing models with up to 78B parameters. For GUI grounding applications, it sets a new standard with 56.1 on OSWorld-G, even outperforming specialized models such as UI-TARS. Our training combines four-stage pre-training (2.4 trillion tokens) with Mixed On-policy Reinforcement Learning (MORL) integrating diverse reward signals. We identify the importance of incorporating high-quality reasoning data with long Chain-of-Thought into pre-training stages, and the benefits of mixed RL despite challenges in simultaneous multi-domain optimization. We also contribute a comprehensive evaluation suite covering 50+ tasks to promote reproducibility and advance the field. The model checkpoints and full evaluation suite are available at https://github.com/XiaomiMiMo/MiMo-VL.