Staff View: :: Library Catalog

Saved in:

Bibliographic Details
Main Authors:	Ling Team, Li, Ang, Liu, Ben, Hu, Binbin, Li, Bing, Zeng, Bingwei, Ye, Borui, Tang, Caizhi, Tian, Changxin, Huang, Chao, Zhang, Chao, Qian, Chen, Ju, Chenchen, Li, Chenchen, Tang, Chengfu, Fu, Chilin, Ren, Chunshao, Wu, Chunwei, Zhang, Cong, Peng, Cunyin, Xu, Dafeng, Wang, Daixin, Zhang, Dalong, Jin, Dingnan, Zhu, Dingyuan, Hu, Dongke, Zhao, Fangzheng, Wu, Feifan, Zhu, Feng, Wang, Gangshan, Zhang, Haitao, Zhao, Hailin, Zhang, Hanxiao, Wang, Hanzi, Qian, Hao, Yu, Haoyi, Zhang, Heng, Zhang, Hongliang, Luan, Hongzhi, Dong, Huirong, Li, Huizhong, Li, Jia, Liu, Jia, Zhu, Jialong, Sha, Jian, Wei, Jianping, Yang, Jiaolong, Ma, Jieyue, Wu, Jiewei, Huang, Jinjing, Tian, Jingyun, Zhang, Jingyuan, Sun, Jinquan, Tu, Juanhui, Liu, Jun, Xu, Jun, Zhou, Jun, Ou, Junjie, Fang, Junpeng, Zhang, Kaihong, Hu, Kaiqin, Shi, Ke, Tang, Kun, Chen, Kunlong, Mei, Lanyin, Liang, Lei, Xu, Lei, Zhang, Libo, Ju, Lin, Yuan, Lin, Zhong, Ling, Ma, Lintao, Liu, Lu, Yu, Lu, Cai, Lun, Zhu, Meiqi, Li, Mengying, Chen, Min, Xue, Minghao, Cai, Minghong, Yin, Mingming, Jiang, Peijie, Zhao, Peilong, Liu, Pingping, Zhao, Qian, Cui, Qing, Huang, Qingxiang, Yang, Qingyuan, Yu, Quankun, Wei, Shaowei, Lian, Shijie, Zheng, Shoujian, Song, Shun, Zhang, Shungen, Zhang, Shuo, Li, Siyuan, Liu, Song, Guo, Ting, Zhao, Tong, Gu, Wanli, Wu, Weichang, Han, Weiguang, Fang, Wenjing, Wang, Wubin, Shu, Xiang, Shi, Xiao, Lan, Xiaoshun, Zhang, Xiaolu, Sun, Xiaqing, Zhao, Xin, Lu, Xingyu, Xu, Xiong, Wang, Xudong, Yang, Xuemin, Yang, Yajie, Xiang, Yang, Li, Yanzhe, Zhang, Yi, Wang, Yilong, Li, Yingxue, Guo, Yongzhen, Fu, Yuzhuo, Wang, Yuanyuan, Yang, Yue, Yu, Yue, Deng, Yufeng, Zhang, Yun, Yu, Yunfei, Zhang, Yuqi, He, Yuxiao, Gui, Zengke, Huan, Zhaoxin, Wang, Zhaoyang, Zhu, Zhibo, Wang, Zhihao, Zhang, Zhiqiang, Wang, Zhoufei, Zeng, Zihang, Liu, Ziqi, Xuan, Zitao, Tang, Zuoli
Format:	Preprint
Published:	2025
Subjects:	Computation and Language Artificial Intelligence
Online Access:	https://arxiv.org/abs/2510.22115
Tags:	Add Tag No Tags, Be the first to tag this record!

_version_	1866912692777779200
author	Ling Team Li, Ang Liu, Ben Hu, Binbin Li, Bing Zeng, Bingwei Ye, Borui Tang, Caizhi Tian, Changxin Huang, Chao Zhang, Chao Qian, Chen Ju, Chenchen Li, Chenchen Tang, Chengfu Fu, Chilin Ren, Chunshao Wu, Chunwei Zhang, Cong Peng, Cunyin Xu, Dafeng Wang, Daixin Zhang, Dalong Jin, Dingnan Zhu, Dingyuan Hu, Dongke Zhao, Fangzheng Wu, Feifan Zhu, Feng Wang, Gangshan Zhang, Haitao Zhao, Hailin Zhang, Hanxiao Wang, Hanzi Qian, Hao Yu, Haoyi Zhang, Heng Zhang, Hongliang Luan, Hongzhi Dong, Huirong Li, Huizhong Li, Jia Liu, Jia Zhu, Jialong Sha, Jian Wei, Jianping Yang, Jiaolong Ma, Jieyue Wu, Jiewei Huang, Jinjing Tian, Jingyun Zhang, Jingyuan Sun, Jinquan Tu, Juanhui Liu, Jun Xu, Jun Zhou, Jun Ou, Junjie Fang, Junpeng Zhang, Kaihong Hu, Kaiqin Shi, Ke Tang, Kun Chen, Kunlong Mei, Lanyin Liang, Lei Xu, Lei Zhang, Libo Ju, Lin Yuan, Lin Zhong, Ling Ma, Lintao Liu, Lu Yu, Lu Cai, Lun Zhu, Meiqi Li, Mengying Chen, Min Xue, Minghao Cai, Minghong Yin, Mingming Jiang, Peijie Zhao, Peilong Liu, Pingping Zhao, Qian Cui, Qing Huang, Qingxiang Yang, Qingyuan Yu, Quankun Wei, Shaowei Lian, Shijie Zheng, Shoujian Song, Shun Zhang, Shungen Zhang, Shuo Li, Siyuan Liu, Song Guo, Ting Zhao, Tong Gu, Wanli Wu, Weichang Han, Weiguang Fang, Wenjing Wang, Wubin Shu, Xiang Shi, Xiao Lan, Xiaoshun Zhang, Xiaolu Sun, Xiaqing Zhao, Xin Lu, Xingyu Xu, Xiong Wang, Xudong Wang, Xudong Yang, Xuemin Yang, Yajie Xiang, Yang Li, Yanzhe Zhang, Yi Wang, Yilong Li, Yingxue Guo, Yongzhen Fu, Yuzhuo Wang, Yuanyuan Yang, Yue Yu, Yue Deng, Yufeng Zhang, Yun Yu, Yunfei Zhang, Yuqi He, Yuxiao Gui, Zengke Huan, Zhaoxin Wang, Zhaoyang Zhu, Zhibo Wang, Zhihao Zhang, Zhiqiang Wang, Zhoufei Zeng, Zihang Liu, Ziqi Xuan, Zitao Tang, Zuoli
author_facet	Ling Team Li, Ang Liu, Ben Hu, Binbin Li, Bing Zeng, Bingwei Ye, Borui Tang, Caizhi Tian, Changxin Huang, Chao Zhang, Chao Qian, Chen Ju, Chenchen Li, Chenchen Tang, Chengfu Fu, Chilin Ren, Chunshao Wu, Chunwei Zhang, Cong Peng, Cunyin Xu, Dafeng Wang, Daixin Zhang, Dalong Jin, Dingnan Zhu, Dingyuan Hu, Dongke Zhao, Fangzheng Wu, Feifan Zhu, Feng Wang, Gangshan Zhang, Haitao Zhao, Hailin Zhang, Hanxiao Wang, Hanzi Qian, Hao Yu, Haoyi Zhang, Heng Zhang, Hongliang Luan, Hongzhi Dong, Huirong Li, Huizhong Li, Jia Liu, Jia Zhu, Jialong Sha, Jian Wei, Jianping Yang, Jiaolong Ma, Jieyue Wu, Jiewei Huang, Jinjing Tian, Jingyun Zhang, Jingyuan Sun, Jinquan Tu, Juanhui Liu, Jun Xu, Jun Zhou, Jun Ou, Junjie Fang, Junpeng Zhang, Kaihong Hu, Kaiqin Shi, Ke Tang, Kun Chen, Kunlong Mei, Lanyin Liang, Lei Xu, Lei Zhang, Libo Ju, Lin Yuan, Lin Zhong, Ling Ma, Lintao Liu, Lu Yu, Lu Cai, Lun Zhu, Meiqi Li, Mengying Chen, Min Xue, Minghao Cai, Minghong Yin, Mingming Jiang, Peijie Zhao, Peilong Liu, Pingping Zhao, Qian Cui, Qing Huang, Qingxiang Yang, Qingyuan Yu, Quankun Wei, Shaowei Lian, Shijie Zheng, Shoujian Song, Shun Zhang, Shungen Zhang, Shuo Li, Siyuan Liu, Song Guo, Ting Zhao, Tong Gu, Wanli Wu, Weichang Han, Weiguang Fang, Wenjing Wang, Wubin Shu, Xiang Shi, Xiao Lan, Xiaoshun Zhang, Xiaolu Sun, Xiaqing Zhao, Xin Lu, Xingyu Xu, Xiong Wang, Xudong Wang, Xudong Yang, Xuemin Yang, Yajie Xiang, Yang Li, Yanzhe Zhang, Yi Wang, Yilong Li, Yingxue Guo, Yongzhen Fu, Yuzhuo Wang, Yuanyuan Yang, Yue Yu, Yue Deng, Yufeng Zhang, Yun Yu, Yunfei Zhang, Yuqi He, Yuxiao Gui, Zengke Huan, Zhaoxin Wang, Zhaoyang Zhu, Zhibo Wang, Zhihao Zhang, Zhiqiang Wang, Zhoufei Zeng, Zihang Liu, Ziqi Xuan, Zitao Tang, Zuoli
contents	We introduce Ling 2.0, a series reasoning-oriented language foundation built upon the principle that every activation boosts reasoning capability. Designed to scale from tens of billions to one trillion parameters under a unified Mixture-of-Experts (MoE) paradigm, Ling 2.0 emphasizes high sparsity, cross-scale consistency, and efficiency guided by empirical scaling laws. The series includes three non-thinking (instruct) models - Ling-mini-2.0, Ling-flash-2.0, and Ling-1T - ranging from 16B to 1T total parameters and achieving up to 7-fold active-compute efficiency compared with dense counterparts. Ling 2.0 integrates coordinated innovations across model architecture, pre-training, post-training, and infrastructure: a high-sparsity MoE with MTP for efficient reasoning, reasoning-oriented data and mid-training CoT activation, reinforcement-based fine-tuning (DFT, Evo-CoT), and full-scale FP8 training with fine-grained heterogeneous pipelines. At the trillion scale, Ling-1T establishes a new Pareto frontier of reasoning accuracy versus computational efficiency, demonstrating that sparse activation, when properly aligned with reasoning objectives, enables scalable and efficient intelligence. Collectively, Ling 2.0 provides a coherent, open, and efficient foundation for advancing future reasoning and thinking models, including the Ring series built upon the same base.
format	Preprint
id	arxiv_https___arxiv_org_abs_2510_22115
institution	arXiv
publishDate	2025
record_format	arxiv
spellingShingle	Every Activation Boosted: Scaling General Reasoner to 1 Trillion Open Language Foundation Ling Team Li, Ang Liu, Ben Hu, Binbin Li, Bing Zeng, Bingwei Ye, Borui Tang, Caizhi Tian, Changxin Huang, Chao Zhang, Chao Qian, Chen Ju, Chenchen Li, Chenchen Tang, Chengfu Fu, Chilin Ren, Chunshao Wu, Chunwei Zhang, Cong Peng, Cunyin Xu, Dafeng Wang, Daixin Zhang, Dalong Jin, Dingnan Zhu, Dingyuan Hu, Dongke Zhao, Fangzheng Wu, Feifan Zhu, Feng Wang, Gangshan Zhang, Haitao Zhao, Hailin Zhang, Hanxiao Wang, Hanzi Qian, Hao Yu, Haoyi Zhang, Heng Zhang, Hongliang Luan, Hongzhi Dong, Huirong Li, Huizhong Li, Jia Liu, Jia Zhu, Jialong Sha, Jian Wei, Jianping Yang, Jiaolong Ma, Jieyue Wu, Jiewei Huang, Jinjing Tian, Jingyun Zhang, Jingyuan Sun, Jinquan Tu, Juanhui Liu, Jun Xu, Jun Zhou, Jun Ou, Junjie Fang, Junpeng Zhang, Kaihong Hu, Kaiqin Shi, Ke Tang, Kun Chen, Kunlong Mei, Lanyin Liang, Lei Xu, Lei Zhang, Libo Ju, Lin Yuan, Lin Zhong, Ling Ma, Lintao Liu, Lu Yu, Lu Cai, Lun Zhu, Meiqi Li, Mengying Chen, Min Xue, Minghao Cai, Minghong Yin, Mingming Jiang, Peijie Zhao, Peilong Liu, Pingping Zhao, Qian Cui, Qing Huang, Qingxiang Yang, Qingyuan Yu, Quankun Wei, Shaowei Lian, Shijie Zheng, Shoujian Song, Shun Zhang, Shungen Zhang, Shuo Li, Siyuan Liu, Song Guo, Ting Zhao, Tong Gu, Wanli Wu, Weichang Han, Weiguang Fang, Wenjing Wang, Wubin Shu, Xiang Shi, Xiao Lan, Xiaoshun Zhang, Xiaolu Sun, Xiaqing Zhao, Xin Lu, Xingyu Xu, Xiong Wang, Xudong Wang, Xudong Yang, Xuemin Yang, Yajie Xiang, Yang Li, Yanzhe Zhang, Yi Wang, Yilong Li, Yingxue Guo, Yongzhen Fu, Yuzhuo Wang, Yuanyuan Yang, Yue Yu, Yue Deng, Yufeng Zhang, Yun Yu, Yunfei Zhang, Yuqi He, Yuxiao Gui, Zengke Huan, Zhaoxin Wang, Zhaoyang Zhu, Zhibo Wang, Zhihao Zhang, Zhiqiang Wang, Zhoufei Zeng, Zihang Liu, Ziqi Xuan, Zitao Tang, Zuoli Computation and Language Artificial Intelligence We introduce Ling 2.0, a series reasoning-oriented language foundation built upon the principle that every activation boosts reasoning capability. Designed to scale from tens of billions to one trillion parameters under a unified Mixture-of-Experts (MoE) paradigm, Ling 2.0 emphasizes high sparsity, cross-scale consistency, and efficiency guided by empirical scaling laws. The series includes three non-thinking (instruct) models - Ling-mini-2.0, Ling-flash-2.0, and Ling-1T - ranging from 16B to 1T total parameters and achieving up to 7-fold active-compute efficiency compared with dense counterparts. Ling 2.0 integrates coordinated innovations across model architecture, pre-training, post-training, and infrastructure: a high-sparsity MoE with MTP for efficient reasoning, reasoning-oriented data and mid-training CoT activation, reinforcement-based fine-tuning (DFT, Evo-CoT), and full-scale FP8 training with fine-grained heterogeneous pipelines. At the trillion scale, Ling-1T establishes a new Pareto frontier of reasoning accuracy versus computational efficiency, demonstrating that sparse activation, when properly aligned with reasoning objectives, enables scalable and efficient intelligence. Collectively, Ling 2.0 provides a coherent, open, and efficient foundation for advancing future reasoning and thinking models, including the Ring series built upon the same base.
title	Every Activation Boosted: Scaling General Reasoner to 1 Trillion Open Language Foundation
topic	Computation and Language Artificial Intelligence
url	https://arxiv.org/abs/2510.22115

Similar Items