GLM-5.2
Open weightsGeneral LLMReasoningTool use
Zhipu AI's flagship GLM model with million-token context window and tool use capabilities for long-horizon coding agents.
1Mcontext
$1.2 / $4.18per 1M tokens (in / out)
753Bparameters
160stories in our feed
Specs
hover a value to see its sourceContext window1M
Max output128K
Parameters753B
Knowledge cutoff—
Licencemit
Commercial useCommercial OK
InputText
OutputText
HF downloads (30d)318.5K
Familyglm
API price
USD per 1M tokensInput$1.2
Output$4.18
Cached input$0.299
Change history
2026-10-21Price $1.4/$4.4 → $1.2/$4.18models.dev
2026-10-13
2026-10-06
2026-10-05
Reflection AI 开源 Beam:501B 参数 MoE,号称最强西方开源模型rohanpaul_ai
2026-10-03
Harness scaling for RL trainingshao__meng
2026-09-29
2026-09-16
2026-09-09
Agent 通过率提升 19%,模型参数未动shao__meng
2026-09-05
Meta参加AIRA₃竞赛模型组合成绩Meta AI
2026-09-03
Trace as State for Long-Context TransformersarXiv: DeepSeek
2026-08-31
GLM-5.3 通过微调达到开放模型最高分DeepLearning.AI
2026-08-29
GLM 5.3 在 Fireworks 上线Fireworks AI
2026-08-18
Qwen 3.8 27B拿下52分,追平GPT-5.6 LunaSimon Willison’s Weblog
2026-08-15
GLM 5.3与5.2同基座 后训练带来大幅提升orange.ai
2026-08-14
GLM-5.3 即将登陆 LMArena,对比 GLM-5.1 与 5.2lmarena.ai
GLM-5.3正式发布,743B模型后训练实现能力跃升shao__meng
GLM-5.3发布,主打安全与长程任务执行歸藏(guizang.ai)
2026-08-05
2026-08-03
GLM-5.2 (Max) Code Arena前端第2,开源模型第一lmarena.ai
2026-07-31
2026-07-30
depthfirstlabs 发布 dfs-large1,专注企业代码库漏洞挖掘Fireworks AI
2026-07-28
Hugging Face 被入侵后靠开源模型 GLM 5.2 自救,Perplexity 加入 OSAAAravind Srinivas
2026-07-27
2026-07-24
HuggingFace入侵者竟是AI自主智能体,GLM-5.2参与分析AI Engineer
2026-07-23
2026-07-19
阿里云 Qwen 3.8 max preview 开源权重模型预览上线,2.4T参数Paul Couvert
2026-07-17
2026-07-16
SGLang在8xB300上以500 TPS服务GLM5.2 NVFP4 Agentic工作负载LMSYS Org (SGLang)
2026-07-09
Perplexity发布GLM 5.2编排器模型研究预览,成本为Opus 0.344倍Aravind Srinivas
2026-07-08
2026-07-06