Buchumschlag

Intern-S1: A Scientific Multimodal Foundation Model

Fuente: arXiv

Gespeichert in:

Bibliographische Detailangaben
Hauptverfasser:	Bai, Lei, Cai, Zhongrui, Cao, Yuhang, Cao, Maosong, Cao, Weihan, Chen, Chiyu, Chen, Haojiong, Chen, Kai, Chen, Pengcheng, Chen, Ying, Chen, Yongkang, Cheng, Yu, Chu, Pei, Chu, Tao, Cui, Erfei, Cui, Ganqu, Cui, Long, Cui, Ziyun, Deng, Nianchen, Ding, Ning, Dong, Nanqing, Dong, Peijie, Dou, Shihan, Du, Sinan, Duan, Haodong, Fan, Caihua, Gao, Ben, Gao, Changjiang, Gao, Jianfei, Gao, Songyang, Gao, Yang, Gao, Zhangwei, Ge, Jiaye, Ge, Qiming, Gu, Lixin, Gu, Yuzhe, Guo, Aijia, Guo, Qipeng, Guo, Xu, He, Conghui, He, Junjun, Hong, Yili, Hou, Siyuan, Hu, Caiyu, Hu, Hanglei, Hu, Jucheng, Hu, Ming, Hua, Zhouqi, Huang, Haian, Huang, Junhao, Huang, Xu, Huang, Zixian, Jiang, Zhe, Kong, Lingkai, Li, Linyang, Li, Peiji, Li, Pengze, Li, Shuaibin, Li, Tianbin, Li, Wei, Li, Yuqiang, Lin, Dahua, Lin, Junyao, Lin, Tianyi, Lin, Zhishan, Liu, Hongwei, Liu, Jiangning, Liu, Jiyao, Liu, Junnan, Liu, Kai, Liu, Kaiwen, Liu, Kuikun, Liu, Shichun, Liu, Shudong, Liu, Wei, Liu, Xinyao, Liu, Yuhong, Liu, Zhan, Lu, Yinquan, Lv, Haijun, Lv, Hongxia, Lv, Huijie, Lv, Qitan, Lv, Ying, Lyu, Chengqi, Ma, Chenglong, Ma, Jianpeng, Ma, Ren, Ma, Runmin, Ma, Runyuan, Ma, Xinzhu, Ma, Yichuan, Ma, Zihan, Mi, Sixuan, Ning, Junzhi, Ning, Wenchang, Pang, Xinle, Peng, Jiahui, Peng, Runyu, Qiao, Yu, Qiu, Jiantao, Qu, Xiaoye, Qu, Yuan, Ren, Yuchen, Shang, Fukai, Shao, Wenqi, Shen, Junhao, Shen, Shuaike, Song, Chunfeng, Song, Demin, Song, Diping, Su, Chenlin, Su, Weijie, Sun, Weigao, Sun, Yu, Tan, Qian, Tang, Cheng, Tang, Huanze, Tang, Kexian, Tang, Shixiang, Tong, Jian, Wang, Aoran, Wang, Bin, Wang, Dong, Wang, Lintao, Wang, Rui, Wang, Weiyun, Wang, Wenhai, Wang, Jiaqi, Wang, Yi, Wang, Ziyi, Wu, Ling-I, Wu, Wen, Wu, Yue, Wu, Zijian, Xiao, Linchen, Xing, Shuhao, Xu, Chao, Xu, Huihui, Xu, Jun, Xu, Ruiliang, Xu, Wanghan, Yang, GanLin, Yang, Yuming, Ye, Haochen, Ye, Jin, Ye, Shenglong, Yu, Jia, Yu, Jiashuo, Yu, Jing, Yuan, Fei, Zang, Yuhang, Zhang, Bo, Zhang, Chao, Zhang, Chen, Zhang, Hongjie, Zhang, Jin, Zhang, Qiaosheng, Zhang, Qiuyinzhe, Zhang, Songyang, Zhang, Taolin, Zhang, Wenlong, Zhang, Wenwei, Zhang, Yechen, Zhang, Ziyang, Zhao, Haiteng, Zhao, Qian, Zhao, Xiangyu, Zhou, Bowen, Zhou, Dongzhan, Zhou, Peiheng, Zhou, Yuhao, Zhou, Yunhua, Zhu, Dongsheng, Zhu, Lin, Zou, Yicheng
Format:	Preprint
Veröffentlicht:	2025
Schlagworte:	Machine Learning Computation and Language Computer Vision and Pattern Recognition
Online-Zugang:	Acceder al recurso
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!

_version_	1866912551139278848
author	Bai, Lei Cai, Zhongrui Cao, Yuhang Cao, Maosong Cao, Weihan Chen, Chiyu Chen, Haojiong Chen, Kai Chen, Pengcheng Chen, Ying Chen, Yongkang Cheng, Yu Chu, Pei Chu, Tao Cui, Erfei Cui, Ganqu Cui, Long Cui, Ziyun Deng, Nianchen Ding, Ning Dong, Nanqing Dong, Peijie Dou, Shihan Du, Sinan Duan, Haodong Fan, Caihua Gao, Ben Gao, Changjiang Gao, Jianfei Gao, Songyang Gao, Yang Gao, Zhangwei Ge, Jiaye Ge, Qiming Gu, Lixin Gu, Yuzhe Guo, Aijia Guo, Qipeng Guo, Xu He, Conghui He, Junjun Hong, Yili Hou, Siyuan Hu, Caiyu Hu, Hanglei Hu, Jucheng Hu, Ming Hua, Zhouqi Huang, Haian Huang, Junhao Huang, Xu Huang, Zixian Jiang, Zhe Kong, Lingkai Li, Linyang Li, Peiji Li, Pengze Li, Shuaibin Li, Tianbin Li, Wei Li, Yuqiang Lin, Dahua Lin, Junyao Lin, Tianyi Lin, Zhishan Liu, Hongwei Liu, Jiangning Liu, Jiyao Liu, Junnan Liu, Kai Liu, Kaiwen Liu, Kuikun Liu, Shichun Liu, Shudong Liu, Wei Liu, Xinyao Liu, Yuhong Liu, Zhan Lu, Yinquan Lv, Haijun Lv, Hongxia Lv, Huijie Lv, Qitan Lv, Ying Lyu, Chengqi Ma, Chenglong Ma, Jianpeng Ma, Ren Ma, Runmin Ma, Runyuan Ma, Xinzhu Ma, Yichuan Ma, Zihan Mi, Sixuan Ning, Junzhi Ning, Wenchang Pang, Xinle Peng, Jiahui Peng, Runyu Qiao, Yu Qiu, Jiantao Qu, Xiaoye Qu, Yuan Ren, Yuchen Shang, Fukai Shao, Wenqi Shen, Junhao Shen, Shuaike Song, Chunfeng Song, Demin Song, Diping Su, Chenlin Su, Weijie Sun, Weigao Sun, Yu Tan, Qian Tang, Cheng Tang, Huanze Tang, Kexian Tang, Shixiang Tong, Jian Wang, Aoran Wang, Bin Wang, Dong Wang, Lintao Wang, Rui Wang, Weiyun Wang, Wenhai Wang, Jiaqi Wang, Yi Wang, Ziyi Wu, Ling-I Wu, Wen Wu, Yue Wu, Zijian Xiao, Linchen Xing, Shuhao Xu, Chao Xu, Huihui Xu, Jun Xu, Ruiliang Xu, Wanghan Yang, GanLin Yang, Yuming Ye, Haochen Ye, Jin Ye, Shenglong Yu, Jia Yu, Jiashuo Yu, Jing Yuan, Fei Zang, Yuhang Zhang, Bo Zhang, Chao Zhang, Chen Zhang, Hongjie Zhang, Jin Zhang, Qiaosheng Zhang, Qiuyinzhe Zhang, Songyang Zhang, Taolin Zhang, Wenlong Zhang, Wenwei Zhang, Yechen Zhang, Ziyang Zhao, Haiteng Zhao, Qian Zhao, Xiangyu Zhao, Xiangyu Zhou, Bowen Zhou, Dongzhan Zhou, Peiheng Zhou, Yuhao Zhou, Yunhua Zhu, Dongsheng Zhu, Lin Zou, Yicheng
author_facet	Bai, Lei Cai, Zhongrui Cao, Yuhang Cao, Maosong Cao, Weihan Chen, Chiyu Chen, Haojiong Chen, Kai Chen, Pengcheng Chen, Ying Chen, Yongkang Cheng, Yu Chu, Pei Chu, Tao Cui, Erfei Cui, Ganqu Cui, Long Cui, Ziyun Deng, Nianchen Ding, Ning Dong, Nanqing Dong, Peijie Dou, Shihan Du, Sinan Duan, Haodong Fan, Caihua Gao, Ben Gao, Changjiang Gao, Jianfei Gao, Songyang Gao, Yang Gao, Zhangwei Ge, Jiaye Ge, Qiming Gu, Lixin Gu, Yuzhe Guo, Aijia Guo, Qipeng Guo, Xu He, Conghui He, Junjun Hong, Yili Hou, Siyuan Hu, Caiyu Hu, Hanglei Hu, Jucheng Hu, Ming Hua, Zhouqi Huang, Haian Huang, Junhao Huang, Xu Huang, Zixian Jiang, Zhe Kong, Lingkai Li, Linyang Li, Peiji Li, Pengze Li, Shuaibin Li, Tianbin Li, Wei Li, Yuqiang Lin, Dahua Lin, Junyao Lin, Tianyi Lin, Zhishan Liu, Hongwei Liu, Jiangning Liu, Jiyao Liu, Junnan Liu, Kai Liu, Kaiwen Liu, Kuikun Liu, Shichun Liu, Shudong Liu, Wei Liu, Xinyao Liu, Yuhong Liu, Zhan Lu, Yinquan Lv, Haijun Lv, Hongxia Lv, Huijie Lv, Qitan Lv, Ying Lyu, Chengqi Ma, Chenglong Ma, Jianpeng Ma, Ren Ma, Runmin Ma, Runyuan Ma, Xinzhu Ma, Yichuan Ma, Zihan Mi, Sixuan Ning, Junzhi Ning, Wenchang Pang, Xinle Peng, Jiahui Peng, Runyu Qiao, Yu Qiu, Jiantao Qu, Xiaoye Qu, Yuan Ren, Yuchen Shang, Fukai Shao, Wenqi Shen, Junhao Shen, Shuaike Song, Chunfeng Song, Demin Song, Diping Su, Chenlin Su, Weijie Sun, Weigao Sun, Yu Tan, Qian Tang, Cheng Tang, Huanze Tang, Kexian Tang, Shixiang Tong, Jian Wang, Aoran Wang, Bin Wang, Dong Wang, Lintao Wang, Rui Wang, Weiyun Wang, Wenhai Wang, Jiaqi Wang, Yi Wang, Ziyi Wu, Ling-I Wu, Wen Wu, Yue Wu, Zijian Xiao, Linchen Xing, Shuhao Xu, Chao Xu, Huihui Xu, Jun Xu, Ruiliang Xu, Wanghan Yang, GanLin Yang, Yuming Ye, Haochen Ye, Jin Ye, Shenglong Yu, Jia Yu, Jiashuo Yu, Jing Yuan, Fei Zang, Yuhang Zhang, Bo Zhang, Chao Zhang, Chen Zhang, Hongjie Zhang, Jin Zhang, Qiaosheng Zhang, Qiuyinzhe Zhang, Songyang Zhang, Taolin Zhang, Wenlong Zhang, Wenwei Zhang, Yechen Zhang, Ziyang Zhao, Haiteng Zhao, Qian Zhao, Xiangyu Zhao, Xiangyu Zhou, Bowen Zhou, Dongzhan Zhou, Peiheng Zhou, Yuhao Zhou, Yunhua Zhu, Dongsheng Zhu, Lin Zou, Yicheng
contents	In recent years, a plethora of open-source foundation models have emerged, achieving remarkable progress in some widely attended fields, with performance being quite close to that of closed-source models. However, in high-value but more challenging scientific professional fields, either the fields still rely on expert models, or the progress of general foundation models lags significantly compared to those in popular areas, far from sufficient for transforming scientific research and leaving substantial gap between open-source models and closed-source models in these scientific domains. To mitigate this gap and explore a step further toward Artificial General Intelligence (AGI), we introduce Intern-S1, a specialized generalist equipped with general understanding and reasoning capabilities with expertise to analyze multiple science modal data. Intern-S1 is a multimodal Mixture-of-Experts (MoE) model with 28 billion activated parameters and 241 billion total parameters, continually pre-trained on 5T tokens, including over 2.5T tokens from scientific domains. In the post-training stage, Intern-S1 undergoes offline and then online reinforcement learning (RL) in InternBootCamp, where we propose Mixture-of-Rewards (MoR) to synergize the RL training on more than 1000 tasks simultaneously. Through integrated innovations in algorithms, data, and training systems, Intern-S1 achieved top-tier performance in online RL training. On comprehensive evaluation benchmarks, Intern-S1 demonstrates competitive performance on general reasoning tasks among open-source models and significantly outperforms open-source models in scientific domains, surpassing closed-source state-of-the-art models in professional tasks, such as molecular synthesis planning, reaction condition prediction, predicting thermodynamic stabilities for crystals. Our models are available at https://huggingface.co/internlm/Intern-S1.
format	Preprint
id	arxiv_https___arxiv_org_abs_2508_15763
institution	arXiv
publishDate	2025
record_format	arxiv
spellingShingle	Intern-S1: A Scientific Multimodal Foundation Model Bai, Lei Cai, Zhongrui Cao, Yuhang Cao, Maosong Cao, Weihan Chen, Chiyu Chen, Haojiong Chen, Kai Chen, Pengcheng Chen, Ying Chen, Yongkang Cheng, Yu Chu, Pei Chu, Tao Cui, Erfei Cui, Ganqu Cui, Long Cui, Ziyun Deng, Nianchen Ding, Ning Dong, Nanqing Dong, Peijie Dou, Shihan Du, Sinan Duan, Haodong Fan, Caihua Gao, Ben Gao, Changjiang Gao, Jianfei Gao, Songyang Gao, Yang Gao, Zhangwei Ge, Jiaye Ge, Qiming Gu, Lixin Gu, Yuzhe Guo, Aijia Guo, Qipeng Guo, Xu He, Conghui He, Junjun Hong, Yili Hou, Siyuan Hu, Caiyu Hu, Hanglei Hu, Jucheng Hu, Ming Hua, Zhouqi Huang, Haian Huang, Junhao Huang, Xu Huang, Zixian Jiang, Zhe Kong, Lingkai Li, Linyang Li, Peiji Li, Pengze Li, Shuaibin Li, Tianbin Li, Wei Li, Yuqiang Lin, Dahua Lin, Junyao Lin, Tianyi Lin, Zhishan Liu, Hongwei Liu, Jiangning Liu, Jiyao Liu, Junnan Liu, Kai Liu, Kaiwen Liu, Kuikun Liu, Shichun Liu, Shudong Liu, Wei Liu, Xinyao Liu, Yuhong Liu, Zhan Lu, Yinquan Lv, Haijun Lv, Hongxia Lv, Huijie Lv, Qitan Lv, Ying Lyu, Chengqi Ma, Chenglong Ma, Jianpeng Ma, Ren Ma, Runmin Ma, Runyuan Ma, Xinzhu Ma, Yichuan Ma, Zihan Mi, Sixuan Ning, Junzhi Ning, Wenchang Pang, Xinle Peng, Jiahui Peng, Runyu Qiao, Yu Qiu, Jiantao Qu, Xiaoye Qu, Yuan Ren, Yuchen Shang, Fukai Shao, Wenqi Shen, Junhao Shen, Shuaike Song, Chunfeng Song, Demin Song, Diping Su, Chenlin Su, Weijie Sun, Weigao Sun, Yu Tan, Qian Tang, Cheng Tang, Huanze Tang, Kexian Tang, Shixiang Tong, Jian Wang, Aoran Wang, Bin Wang, Dong Wang, Lintao Wang, Rui Wang, Weiyun Wang, Wenhai Wang, Jiaqi Wang, Yi Wang, Ziyi Wu, Ling-I Wu, Wen Wu, Yue Wu, Zijian Xiao, Linchen Xing, Shuhao Xu, Chao Xu, Huihui Xu, Jun Xu, Ruiliang Xu, Wanghan Yang, GanLin Yang, Yuming Ye, Haochen Ye, Jin Ye, Shenglong Yu, Jia Yu, Jiashuo Yu, Jing Yuan, Fei Zang, Yuhang Zhang, Bo Zhang, Chao Zhang, Chen Zhang, Hongjie Zhang, Jin Zhang, Qiaosheng Zhang, Qiuyinzhe Zhang, Songyang Zhang, Taolin Zhang, Wenlong Zhang, Wenwei Zhang, Yechen Zhang, Ziyang Zhao, Haiteng Zhao, Qian Zhao, Xiangyu Zhao, Xiangyu Zhou, Bowen Zhou, Dongzhan Zhou, Peiheng Zhou, Yuhao Zhou, Yunhua Zhu, Dongsheng Zhu, Lin Zou, Yicheng Machine Learning Computation and Language Computer Vision and Pattern Recognition In recent years, a plethora of open-source foundation models have emerged, achieving remarkable progress in some widely attended fields, with performance being quite close to that of closed-source models. However, in high-value but more challenging scientific professional fields, either the fields still rely on expert models, or the progress of general foundation models lags significantly compared to those in popular areas, far from sufficient for transforming scientific research and leaving substantial gap between open-source models and closed-source models in these scientific domains. To mitigate this gap and explore a step further toward Artificial General Intelligence (AGI), we introduce Intern-S1, a specialized generalist equipped with general understanding and reasoning capabilities with expertise to analyze multiple science modal data. Intern-S1 is a multimodal Mixture-of-Experts (MoE) model with 28 billion activated parameters and 241 billion total parameters, continually pre-trained on 5T tokens, including over 2.5T tokens from scientific domains. In the post-training stage, Intern-S1 undergoes offline and then online reinforcement learning (RL) in InternBootCamp, where we propose Mixture-of-Rewards (MoR) to synergize the RL training on more than 1000 tasks simultaneously. Through integrated innovations in algorithms, data, and training systems, Intern-S1 achieved top-tier performance in online RL training. On comprehensive evaluation benchmarks, Intern-S1 demonstrates competitive performance on general reasoning tasks among open-source models and significantly outperforms open-source models in scientific domains, surpassing closed-source state-of-the-art models in professional tasks, such as molecular synthesis planning, reaction condition prediction, predicting thermodynamic stabilities for crystals. Our models are available at https://huggingface.co/internlm/Intern-S1.
title	Intern-S1: A Scientific Multimodal Foundation Model
topic	Machine Learning Computation and Language Computer Vision and Pattern Recognition
url	https://arxiv.org/abs/2508.15763