zai-org/GLM-5.2-FP8
摘要
Z.AI 发布 GLM-5.2,一款旗舰级开源模型,拥有可靠的 1M token 上下文窗口,改进的编码能力,以及新的 IndexShare 稀疏注意力架构,在 1M 上下文下 FLOPs 减少了 2.9 倍。
查看缓存全文
缓存时间: 2026/06/18 11:40
zai-org/GLM-5.2-FP8 · Hugging Face
来源:https://huggingface.co/zai-org/GLM-5.2-FP8
👋 加入我们的WeChat (https://raw.githubusercontent.com/zai-org/GLM-5/refs/heads/main/resources/wechat.png)或Discord (https://discord.gg/QR7SARHRxK)社区。📖 查看 GLM-5.2 博客 (https://z.ai/blog/glm-5.2)和 GLM-5 技术报告 (https://arxiv.org/abs/2602.15763)。📍 在 Z.ai API 平台 (https://docs.z.ai/guides/llm/glm-5.2)上使用 GLM-5.2 API 服务。🔜 在此处 (https://chat.z.ai/)尝试 GLM-5.2。
[论文 (https://huggingface.co/papers/2602.15763)] [GitHub (https://github.com/zai-org/GLM-5)]
https://huggingface.co/zai-org/GLM-5.2-FP8#introduction介绍
我们推出了 GLM-5.2,这是我们的最新旗舰模型,专为长时程任务打造。相比前代 GLM-5.1,它在长时程任务能力上实现了大幅飞跃,并且首次在坚实的 100万 Token 上下文上提供了这种能力。GLM-5.2 的新能力包括:
- 100万 Token 稳定上下文:一个坚实的 100万 Token 上下文,稳定支撑长时程工作
- 高级编码与灵活推理深度:更强的编码能力,支持多种推理深度级别以平衡性能与延迟
- 改进架构:我们提出了 IndexShare (https://arxiv.org/abs/2603.12201),在每四个稀疏注意力层中复用同一个索引器,在 100万 上下文长度下将每个 Token 的 FLOPs 降低 2.9 倍。我们还改进了 GLM-5.2 的 MTP 层用于推测解码,将接受长度最多提升 20%
- 完全开源:采用 MIT 开源许可证——无地域限制,无边界的技术访问
bench_52 (https://raw.githubusercontent.com/zai-org/GLM-5/refs/heads/main/resources/bench_52.png)
https://huggingface.co/zai-org/GLM-5.2-FP8#benchmark基准测试
| 基准测试 | GLM-5.2 | GLM-5.1 | Qwen3.7-Max | MiniMax M3 | DeepSeek-V4-Pro | Claude Opus 4.8 | GPT-5.5 | Gemini 3.1 Pro |
|---|---|---|---|---|---|---|---|---|
| 推理 | ||||||||
| HLE | 40.5 | 31 | 41.4 | 37 | 43 | 7 | 49.8* | 41.4* |
| HLE(使用工具) | 54.7 | 52.3 | 53.5 | - | 48.2 | 57.9* | 52.2* | 51.4* |
| CritPt | 20.9 | 4.6 | 13.4 | 3.7 | 12.9 | 20.9 | 27.1 | 17.7 |
| AIME 2026 | 99.2 | 95.3 | 97 | - | 94.6 | 95.7 | 98.3 | 98.2 |
| HMMT Nov. 2025 | 94.4 | 94 | 95 | 84.4 | 94.4 | 96.5 | 96.5 | 94.8 |
| HMMT Feb. 2026 | 92.5 | 82.5 | 97.1 | 84.4 | 95.2 | 96.7 | 96.7 | 87.3 |
| IMOAnswerBench | 91.0 | 83.8 | 90 | - | 89.8 | 83.5 | - | 81 |
| GPQA-Diamond | 91.2 | 86.2 | 90 | 93 | 90.1 | 93.6 | 93.6 | 94.3 |
| 编码 | ||||||||
| SWE-bench Pro | 62.1 | 58.4 | 60.6 | 59 | 55.4 | 69.2 | 58.6 | 54.2 |
| NL2Repo | 48.9 | 42.7 | 47.2 | 42.1 | 35.5 | 69.7 | 50.7 | 33.4 |
| DeepSWE | 46.2 | 18 | 18 | 20 | 8 | 58 | 70 | 10 |
| ProgramBench | 63.7 | 50.9 | - | - | 47.8 | 71.9 | 70.8 | 39.5 |
| Terminal Bench 2.1 (Terminus-2) | 81.0 | 63.5 | 75 | 65 | 64 | 85 | 84 | 74 |
| Terminal Bench 2.1 (最佳公开包) | 82.7 | 69 | - | - | - | 78.9 | 83.4 | 70.7 |
| FrontierSWE (Dominance) | 74.4 | 30.5 | - | - | 29.0 | 75.1 | 72.6 | 39.6 |
| PostTrainBench | 34.3 | 20.1 | - | - | - | 37.2 | 28.4 | 21.6 |
| SWE-Marathon | 13.0 | 1.0 | - | - | - | 26.0 | 12.0 | 4.0 |
| 智能体 | ||||||||
| MCP-Atlas (公开集) | 76.8 | 71.8 | 76.4 | 74.2 | 73.6 | 77.8 | 75.3 | 69.2 |
| Tool-Decathlon | 48.2 | 40.7 | - | - | 52.8 | 59.9 | 55.6 | 48.8 |
https://huggingface.co/zai-org/GLM-5.2-FP8#serve-glm-52-locally本地运行 GLM-5.2
GLM-5.2 支持使用以下框架部署,欢迎尝试:
- SGLang (https://github.com/sgl-project/sglang)(v0.5.13.post1+)—— 参见 cookbook (https://cookbook.sglang.io/autoregressive/GLM/GLM-5.2)
- vLLM (https://github.com/vllm-project/vllm)(v0.23.0+)—— 参见 recipes (https://recipes.vllm.ai/zai-org/GLM-5.2)
- Transformers (https://github.com/huggingface/transformers)(v0.5.12+)—— 参见 transformers 文档 (https://github.com/huggingface/transformers/blob/main/docs/source/en/model_doc/glm_moe_dsa.md)
- KTransformers (https://github.com/kvcache-ai/ktransformers)(v0.5.12+)—— 参见教程 (https://github.com/kvcache-ai/ktransformers/blob/main/doc/en/kt-kernel/GLM-5.2-Tutorial.md)
- 在
Ascend NPU平台上部署时,支持 vLLM-Ascend、xLLM 和 SGLang 等推理框架 —— 参见此处 (https://huggingface.co/zai-org/GLM-5.2-FP8/blob/main/github.com/zai-org/GLM-5/blob/main/example/ascend.md)。
https://huggingface.co/zai-org/GLM-5.2-FP8#citation引用
如果您在研究中发现 GLM-5.2 有用,请引用我们的技术报告:
@misc{glm5team2026glm5vibecodingagentic, title={GLM-5: from Vibe Coding to Agentic Engineering}, author={GLM-5-Team and : and Aohan Zeng and Xin Lv and Zhenyu Hou and Zhengxiao Du and Qinkai Zheng and Bin Chen and Da Yin and Chendi Ge and Chenghua Huang and Chengxing Xie and Chenzheng Zhu and Congfeng Yin and Cunxiang Wang and Gengzheng Pan and Hao Zeng and Haoke Zhang and Haoran Wang and Huilong Chen and Jiajie Zhang and Jian Jiao and Jiaqi Guo and Jingsen Wang and Jingzhao Du and Jinzhu Wu and Kedong Wang and Lei Li and Lin Fan and Lucen Zhong and Mingdao Liu and Mingming Zhao and Pengfan Du and Qian Dong and Rui Lu and Shuang-Li and Shulin Cao and Song Liu and Ting Jiang and Xiaodong Chen and Xiaohan Zhang and Xuancheng Huang and Xuezhen Dong and Yabo Xu and Yao Wei and Yifan An and Yilin Niu and Yitong Zhu and Yuanhao Wen and Yukuo Cen and Yushi Bai and Zhongpei Qiao and Zihan Wang and Zikang Wang and Zilin Zhu and Ziqiang Liu and Zixuan Li and Bojie Wang and Bosi Wen and Can Huang and Changpeng Cai and Chao Yu and Chen Li and Chengwei Hu and Chenhui Zhang and Dan Zhang and Daoyan Lin and Dayong Yang and Di Wang and Ding Ai and Erle Zhu and Fangzhou Yi and Feiyu Chen and Guohong Wen and Hailong Sun and Haisha Zhao and Haiyi Hu and Hanchen Zhang and Hanrui Liu and Hanyu Zhang and Hao Peng and Hao Tai and Haobo Zhang and He Liu and Hongwei Wang and Hongxi Yan and Hongyu Ge and Huan Liu and Huanpeng Chu and Jia'ni Zhao and Jiachen Wang and Jiajing Zhao and Jiamin Ren and Jiapeng Wang and Jiaxin Zhang and Jiayi Gui and Jiayue Zhao and Jijie Li and Jing An and Jing Li and Jingwei Yuan and Jinhua Du and Jinxin Liu and Junkai Zhi and Junwen Duan and Kaiyue Zhou and Kangjian Wei and Ke Wang and Keyun Luo and Laiqiang Zhang and Leigang Sha and Liang Xu and Lindong Wu and Lintao Ding and Lu Chen and Minghao Li and Nianyi Lin and Pan Ta and Qiang Zou and Rongjun Song and Ruiqi Yang and Shangqing Tu and Shangtong Yang and Shaoxiang Wu and Shengyan Zhang and Shijie Li and Shuang Li and Shuyi Fan and Wei Qin and Wei Tian and Weining Zhang and Wenbo Yu and Wenjie Liang and Xiang Kuang and Xiangmeng Cheng and Xiangyang Li and Xiaoquan Yan and Xiaowei Hu and Xiaoying Ling and Xing Fan and Xingye Xia and Xinyuan Zhang and Xinze Zhang and Xirui Pan and Xu Zou and Xunkai Zhang and Yadi Liu and Yandong Wu and Yanfu Li and Yidong Wang and Yifan Zhu and Yijun Tan and Yilin Zhou and Yiming Pan and Ying Zhang and Yinpei Su and Yipeng Geng and Yong Yan and Yonglin Tan and Yuean Bi and Yuhan Shen and Yuhao Yang and Yujiang Li and Yunan Liu and Yunqing Wang and Yuntao Li and Yurong Wu and Yutao Zhang and Yuxi Duan and Yuxuan Zhang and Zezhen Liu and Zhengtao Jiang and Zhenhe Yan and Zheyu Zhang and Zhixiang Wei and Zhuo Chen and Zhuoer Feng and Zijun Yao and Ziwei Chai and Ziyuan Wang and Zuzhou Zhang and Bin Xu and Minlie Huang and Hongning Wang and Juanzi Li and Yuxiao Dong and Jie Tang}, year={2026}, eprint={2602.15763}, archivePrefix={arXiv}, primaryClass={cs.LG}, url={https://arxiv.org/abs/2602.15763}, }
相似文章
zai-org/GLM-5.2 来了!
Z.AI 发布了 GLM-5.2,这是一款新的旗舰模型,拥有稳定的 1M token 上下文窗口,通过灵活的思考努力增强了编码能力,并通过 IndexShare 改进了架构。该模型在 MIT 开源许可证下发布。
GLM-5.2: 专为长程任务打造
Z.AI推出GLM-5.2,这是一款专为长程任务设计的旗舰模型,拥有稳定的100万token上下文、改进的编码能力以及MIT开源许可证,在与Opus 4.8和GPT-5.5等领先模型的对比中展现了竞争力。
zai-org/GLM-5.1
GLM-5.1 是一款新一代旗舰AI模型,针对代理工程进行了优化,编码能力显著增强,在SWE-Bench Pro上达到了最先进性能,并通过扩展迭代和工具使用展示了卓越的长周期任务处理能力。
GLM-5.2(6分钟阅读)
Z.ai 推出了 GLM-5.2,拥有100万token的上下文窗口、新的推理控制功能,并支持长期编码任务。该模型现已面向 Coding Plan 用户提供,API 访问、聊天机器人支持以及 MIT 许可的开放权重将于下周推出。
GLM-5.2 可能是目前最强大的纯文本开放权重大语言模型
中国AI实验室Z.ai发布了GLM-5.2,这是一个拥有7530亿参数的开放权重大语言模型,支持100万token的上下文窗口,采用MIT许可证。该模型在Artificial Analysis Intelligence Index上获得最高分,并在Code Arena WebDev排行榜上排名第二。