MiniMax-M2系列：微小激活释放最大真实世界智能
MiniMax推出M2系列大语言模型。其旗舰模型M2采用混合专家（MoE）架构，总参数229.9B，每个token仅激活9.8B参数。该系列专为智能体部署设计，基于三大组件构建：智能体驱动的数据管道、可扩展的智能体原生强化学习系统Forge，以及展示早期自我进化能力的M2.7检查点。这种设计使其在智能体编码、深度搜索、办公任务及推理基准测试中达到了前沿性能水平。
MiniMax 把激活参数压到 9.8B 却敢喊前沿，整套 design 都是为 agent 场景重做的，搞 agent 的开发者该认真看一眼这个信号。
Authors:Aili Chen, Aonian Li, Baichuan Zhou, Bangwei Gong, Binyang Jiang, Boji Dan, Changhao Zhang, Changqing Yu, Chao Wang, Cheng Ma, Cheng Zhong, Cheng Zhu, Chengjun Xiao, Chengyi Yang, Chengyu Du, Chenyang Zhang, Chi Zhang, Chuangyi Huang, Chunhao Zhang, Chunhui Du, Chunyu Zhao, Congchao Guo, Da Chen, Deming Ding, Dianjun Sun, Dong Li, Dongyu Zhang, Enhui Yang, Fei Yu, Guang Zheng, Guodong Zheng, Guohong Li, Haichao Zhu, Haigang Zhou, Haimo Zhang, Han Ding, Hao Zhang, Haohai Sun, Haolin Lyu, Haonan Lu, Haoyu Wang, Huajie Shi, Huiyang Li, Jiacheng Chen, Jian Zhang, Jiaqi Zhuang, Jiaren Cai, Jiaxin Pan, Jiayao Li, Jiayuan Song, Jichuan Zhang, Jie Wang, Jihao Gu, Jin Zhu, Jingwei Dong, Jingyang Li, Jingyu Zhang, Jingze Zhuang, Jinhao Tian, Jinli Liu, Jinyi Hu, Jun Tao, Jun Zhang, Junbin Ruan, Junhao Xu, Junjie Yan, Junteng Liu, Junxian He, Kang Xu, Ke Ji, Ke Yang, Kecheng Xiao, Keyu Duan, Keyu Li, Le Han, Letian Ruan, Li Yuan, Lianfei Yu, Liheng Feng, Lijie Mo, Lin Li, Linge Du, Lingye Bao, Lingyu Yang, Lingyuan Zhou, Loki, Lu Chen, Lunbin Zeng, Ming Li, Ming Zhong, Mingliang Tao, Mingyuan Chi, Mujie Lin, Nan Hu, Ningxin Chen, Peiyin Zhu, Peng Gao, Pengcheng Gao, Pengfei Li, Penglin Li
et al. (118 additional authors not shown)
Abstract:We introduce the MiniMax-M2 series, a family of Mixture-of-Experts language models built around the principle that mini activations can unleash maximum real-world intelligence. The flagship M2 contains 229.9B total parameters with only 9.8B activated per token. Designed end-to-end for agentic deployment, the M2 series rests on three components: (i) agent-driven data pipelines producing large-scale, verifiable trajectories across agentic coding and agentic cowork, each grounded in an executable workspace and an artifact-aligned reward; (ii) Forge, a scalable agent-native RL system that adapts to long-horizon agent trajectories, paired with windowed-FIFO scheduling, prefix-tree merging, inference optimization, and a clean training-inference-agent decoupling that supports both white-box and black-box agents; (iii) the latest M2.7 checkpoint takes an early step toward self-evolution -- autonomously debugging training runs and modifying its own scaffold. Across M2 through M2.7, this combination translates a mini-activation footprint into frontier-tier performance on agentic coding, deep search, office-task, and reasoning benchmarks.
| Comments: | Technical Report. 35 pages, 10 figures, 4 tables |
| Subjects: | Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Machine Learning (cs.LG) |
| Cite as: | arXiv:2605.26494 [cs.AI] |
| (or arXiv:2605.26494v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2605.26494 arXiv-issued DOI via DataCite |
Submission history
From: Li Yuan [view email]
[v1]
Tue, 26 May 2026 03:16:11 UTC (1,591 KB)
[v2]
Thu, 30 Jul 2026 05:04:36 UTC (1,593 KB)
来源：HuggingFace Daily Papers（社区热门论文） · arxiv.org