GLM-4.5:智能体、推理和编码(ARC)基础模型

发表
Tiezhen WANGTiezhen WANG 提交
作者: GLM-4. 5 Team, Zeng AohanAohan Zeng, Xin LvXin Lv, Qinkai ZhengQinkai Zheng, Zhenyu Hou, Bin Chen, Chengxing Xie, Cunxiang Wang, Da Yin, Hao Zeng, Jiajie Zhang, Kedong Wang, Lucen Zhong, Mingdao LiuMingdao Liu, Rui LuRui Lu, Shulin Cao, Xiaohan Zhang, Huang XuanchengXuancheng Huang, YaoWeiYao Wei, Yean ChengYean Cheng, Yifan An, Yilin Niu, Yuanhao Wen, Yushi BaiYushi Bai, Zhengxiao Du, Zihan Wang, Zilin Zhu, Bohan Zhang, Bosi Wen, Bowen Wu, Bowen Xu, Can Huang, Casey Zhao, Changpeng Cai, Chao Yu, Chen Li, Chendi Ge, Chenghua Huang, Chenhui Zhang, Chenxi Xu, Chenzheng Zhu, Chuang Li, Congfeng Yin, Daoyan Lin, Dayong Yang, Dazhi JiangDazhi Jiang, Ding Ai, Erle Zhu, Fei Wang, Gengzheng Pan, Guo Wang, hi loongHailong Sun, Haitao Li, Haiyang Li, Haiyi Hu, Hanyu Zhang, Hao Peng, Hao Tai, Haoke Zhang, Haoran Wang, Haoyu Yang, He Liu, He Zhao, Hongwei Liu, Hongxi Yan, Huan Liu, Huilong Chen, Ji Li, Jiajing Zhao, Jiamin Ren, Jian Jiao, Jiani Zhao, Jianyang Yan, Jiaqi Wang, Jiayi Gui, Jiayue Zhao, Jie Liu, Jijie Li, Jing Li, Jing Lu, Jingsen Wang, Jingwei Yuan, david leeJingxuan Li, Jingzhao Du, Jinhua DuJinhua Du, Jinxin Liu, Junkai ZhiJunkai Zhi, Junli Gao, Ke Wang, Yang LekangLekang Yang, Liang Xu, Lin Fan, Lindong Wu, Lintao Ding, Lu Wang, Man Zhang, Minghao Li, Minghuan Xu, Mingming Zhao, Mingshu Zhai, Pengfan Du, Qian DongQian Dong, Shangde Lei, Shangqing TuShangqing Tu, Shangtong Yang, Shaoyou Lu, Shijie Li, Shuang Li, Shuang-Li, Shuxun Yang, Sibo Yi, Tianshu Yu, Wei Tian, Weihan Wang, Wenbo Yu, Weng Lam Tam, Wenjie Liang, Wentao Liu, Xiao WangXiao Wang, Xiaohan Jia, Xiaotao Gu, Xiaoying Ling, Xin Wang, Xing Fan, Xingru Pan, Xinyuan Zhang, Xinze Zhang, Xiuqing Fu, Xunkai Zhang, Yabo Xu, Yandong Wu, Yida Lu, Yidong Wang, Yilin Zhou, Yiming Pan, Ying Zhang, Yingli Wang, Yingru LiYingru Li, Yinpei SuYinpei Su, Yipeng Geng, Yitong Zhu, Yongkun Yang, Yuhang Li, wuyuhaoYuhao Wu, Yujiang Li, Yunan Liu, Yunqing Wang, Yuntao Li, zRYuxuan Zhang, Zezhen Liu, yzZhen Yang, Zhengda Zhou, Zhongpei Qiao, Zhuoer Feng, Zhuorui Liu, Zichen Zhang, Zihan Wang, Zijun Yao, Zikang Wang, Ziqiang LiuZiqiang Liu, Ziwei Chai, Zixuan LiZixuan Li, Zuodong Zhao, Wenguang Chen, Jidong Zhai, Bin Xu, Minlie Huang, Hongning Wang, Juanzi Li, Yuxiao Dong, Jie Tang

摘要

AI 生成总结
GLM-4.5,一个拥有3550亿参数的专家混合大型语言模型,通过多阶段训练和强化学习,在代理、推理和编码任务中取得了强大的性能。
我们推出 GLM-4.5,这是一款开源的专家混合 (MoE) 大型语言模型,总参数量为 3550 亿,激活参数量为 320 亿,采用混合推理方法,支持思考和直接响应两种模式。通过在 23 万亿个 token 上进行多阶段训练,并通过专家模型迭代和强化学习进行全面的后期训练,GLM-4.5 在代理、推理和编码 (ARC) 任务中取得了强大性能,在 TAU-Bench 上得分 70.1%,在 AIME 24 上得分 91.0%,在 SWE-bench Verified 上得分 64.2%。尽管参数量远少于几个竞争对手,GLM-4.5 在所有评估模型中总体排名第 3,在代理基准测试中排名第 2。我们发布 GLM-4.5(3550 亿参数)和紧凑版本 GLM-4.5-Air(1060 亿参数),以推动推理和代理 AI 系统领域的研究。代码、模型和更多信息可在 https://github.com/zai-org/GLM-4.5 获取。
查看 arXiv 页面查看 PDF

评论

Tiezhen WANGTiezhen WANG
论文提交者

GLM 4.5 技术论文

MagaMaga

在您的演示之前,我们已经在 Moe 的平台上完成了。不过,谢谢。

LexLex

👍

buyubuyu

4xupscale.png

buyubuyu

你能描述一下这张图片吗?

Rasel AHmedRasel AHmed

画一只老虎

I am GHOSTI am GHOST

简单的 HTML 文件