titleMiMo-V2.6: Scaling Reinforcement Learning Towards Self-Improvement
authorsCore Team, Zongming Qiao, Ziyue Hua, Zirui Ou, Zihao Yue, Zihan Jiang, Zhuo Huang, Zhiyang Chen, Zhixian Zheng, Zhipeng Xu, Zhengrui Ma, Yuyang Hu, Yuhang Dong, Yuechen Zhang, Yudong Wang, Yuanxin Liu, Yixin Yang, Yishuo Cai, Yikai Zhao, Yihan Yan, Yifan Zhang, Yifan Song, Xiyu Wei, Xing Zhang, Xin Zhang, Xiaoqian Liu, Xiaodong Ji, Xiangwei Deng, Xueyu Guo, Wenhan Ma, Weimin Xiong, Weikun Wang, Weiji Zhuang, Shuo Liu, Shuhuai Ren, Shuhao Gu, Shimao Chen, Shijie Cao, Shihua Yu, Shicheng Li, Shengjie Zhou, Shaolei Zhang, Rang Li, Qiying Wang, Qingkai Fang, Qianli Chen, Minzheng Wang, Liwen Wang, Linli Yao, Linghao Zhang, Liangyu Cheng, Liang Zhao, Lei Li, Jinhao Dong, Jinyu Xiang, Jianyu Wei, Jiangshan Duo, Huaqiu Liu, Huanjie Fan, Hongyi Guan, Hongshen Xu, Hao Tian, Hanyu Li, Hailin Zhang, Gang Wang, Fuli Luo, Feng Wei, Dong Zhang, Dawei Zhu, Chiheng Lou, Chenhong He, Chenhao He, Chenghua Liu, Bowen Ye, Bowen Shen, Boshen Xu, Bo Yang, Bingquan Xia, Bangjun Xiao, Baixuan Xu, Zhouxiang Mao, Zhiyang Zhang, Zhixiang Xu, Zhenru Lin, Zhengju Tang, Zhaojun Huang, Yuzhe Weng, Yuxing Xiang, Yuxiao Li, Yuheng Yang, Yuhang Wang, Yuchen Liu, Yuanyuan Tian, Yuanliang Dong, Yu Cheng, Yongzhe He, Yongshun Liang, Yong Wang, Yiyan Wang, Yitian Gong, Yijie Zhang, Yanshu Xin, Xun Zhang, Xingjian Zhao, Wenyu Yang, Wenshan Huang, Wenhao Li, Tingwei Huang, Tianyu Yu, Tianyang Lu, Taoyu Yang, Sinan Du, Shutong Tian, Shulin Du, Shengfan Wang, Shanchuan Fang, Qihao Zhang, Qibin Yang, Qian Yu, Qian Tu, Pengrong Xie, Peipei Wang, Peidian Li, Minkun Guo, Mingchen Shao, Luohan Gao, Lijie Wang, Liang Shi, Kaiqi Chen, Kaiming Liu, Kaifei Wang, Kai Yang, Jinlong Xue, Jiechen Zhang, Jiaxuan Liu, Hongxu An, Hao Peng, Hanglong Lü, Guonan Wang, Feiyu Yang, Fanyu Cao, Fangyue Liu, Fan Cui, Cong Wang, Chun Chen, Chenxu Bai, Chengxuan Zhu, Chenghua Wang, Boyi Zeng
date2026-10-08
venuearXiv (preprint, not peer reviewed)
idarXiv:2610.11959
affiliationXiaomi MiMo
upvotes87 (Hugging Face Daily Papers, 2026-11-01)
linksabs pdf hf
abstractReinforcement learning (RL) is the central training paradigm for advancing large foundation models towards self-improvement. This report introduces the MiMo-V2.6 series, an omni-modal family that pushes the frontier of model intelligence by scaling RL compute. Prior to RL, we conduct mid-training on a broad multimodal corpus to provide ample exploration space, and build a solid infrastructure on the pretrained hybrid-SWA architecture to support subsequent scale-up. We scale RL compute along three dimensions: (1) larger batches and higher throughput, with an asynchronous training that consumes 1,568 samples and 2.7-3.7B tokens per step at context lengths of up to 1M; (2) more diverse and complex environments, spanning code, general, visual, and cyber domains under a mixture of agent harnesses; and (3) more grader compute, via groupwise agentic grading that yields more accurate reward signals for long-horizon tasks and steers the model towards shorter, more token-efficient solutions. To keep training stable at scale, we freeze the MoE router and establish a multi-layer defense against reward hacking. We further build infrastructure for mixed-task agentic RL, including a unified trajectory representation, high-concurrency multi-framework rollout, decoupled control and data planes, and training-inference consistency. We open-source the training dynamics, RL environments, and RL framework to facilitate reproduction and further research on scaled RL and model self-improvement.

cite

@misc{team2026mimo,
  title={MiMo-V2.6: Scaling Reinforcement Learning Towards Self-Improvement},
  author={Core Team and Zongming Qiao and Ziyue Hua and Zirui Ou and Zihao Yue and Zihan Jiang and Zhuo Huang and Zhiyang Chen and Zhixian Zheng and Zhipeng Xu and Zhengrui Ma and Yuyang Hu and Yuhang Dong and Yuechen Zhang and Yudong Wang and Yuanxin Liu and Yixin Yang and Yishuo Cai and Yikai Zhao and Yihan Yan and Yifan Zhang and Yifan Song and Xiyu Wei and Xing Zhang and Xin Zhang and Xiaoqian Liu and Xiaodong Ji and Xiangwei Deng and Xueyu Guo and Wenhan Ma and Weimin Xiong and Weikun Wang and Weiji Zhuang and Shuo Liu and Shuhuai Ren and Shuhao Gu and Shimao Chen and Shijie Cao and Shihua Yu and Shicheng Li and Shengjie Zhou and Shaolei Zhang and Rang Li and Qiying Wang and Qingkai Fang and Qianli Chen and Minzheng Wang and Liwen Wang and Linli Yao and Linghao Zhang and Liangyu Cheng and Liang Zhao and Lei Li and Jinhao Dong and Jinyu Xiang and Jianyu Wei and Jiangshan Duo and Huaqiu Liu and Huanjie Fan and Hongyi Guan and Hongshen Xu and Hao Tian and Hanyu Li and Hailin Zhang and Gang Wang and Fuli Luo and Feng Wei and Dong Zhang and Dawei Zhu and Chiheng Lou and Chenhong He and Chenhao He and Chenghua Liu and Bowen Ye and Bowen Shen and Boshen Xu and Bo Yang and Bingquan Xia and Bangjun Xiao and Baixuan Xu and Zhouxiang Mao and Zhiyang Zhang and Zhixiang Xu and Zhenru Lin and Zhengju Tang and Zhaojun Huang and Yuzhe Weng and Yuxing Xiang and Yuxiao Li and Yuheng Yang and Yuhang Wang and Yuchen Liu and Yuanyuan Tian and Yuanliang Dong and Yu Cheng and Yongzhe He and Yongshun Liang and Yong Wang and Yiyan Wang and Yitian Gong and Yijie Zhang and Yanshu Xin and Xun Zhang and Xingjian Zhao and Wenyu Yang and Wenshan Huang and Wenhao Li and Tingwei Huang and Tianyu Yu and Tianyang Lu and Taoyu Yang and Sinan Du and Shutong Tian and Shulin Du and Shengfan Wang and Shanchuan Fang and Qihao Zhang and Qibin Yang and Qian Yu and Qian Tu and Pengrong Xie and Peipei Wang and Peidian Li and Minkun Guo and Mingchen Shao and Luohan Gao and Lijie Wang and Liang Shi and Kaiqi Chen and Kaiming Liu and Kaifei Wang and Kai Yang and Jinlong Xue and Jiechen Zhang and Jiaxuan Liu and Hongxu An and Hao Peng and Hanglong Lü and Guonan Wang and Feiyu Yang and Fanyu Cao and Fangyue Liu and Fan Cui and Cong Wang and Chun Chen and Chenxu Bai and Chengxuan Zhu and Chenghua Wang and Boyi Zeng},
  year={2026},
  eprint={2610.11959},
  archivePrefix={arXiv},
  url={https://arxiv.org/abs/2610.11959}
}