推广AI 会员在线充值
访问赞助商 Chatgpt/Claude/Grok 会员24h在线自助充值 安全快速 稳定保障
奥创(Ultra)商城会员充值中心正在加载…
推广Chatgpt/Claude/Grok 会员24h在线自助充值 安全快速 稳定保障
奥创(Ultra)商城会员充值中心
推广本页公开展示实时样本、模型覆盖与响应时间。先用 5 元按量跑自己的 Claude Code/Codex 任务，成功后再决定是否长期使用。
ACS Gateway
推广【腾讯云】WorkBuddy 全场景 AI 办公工作台，团队旗舰版低至198元/月
腾讯云
推广人人可用的AI应用创作平台 自然语言交互，全栈应用搭建，伟大创意快速实现;个人版新用户注册送5000积分
阿里云Z.ai 宣布 GLM-5.3 开放权重，定位于智能体编程和网络防御任务。模型权重已在 Hugging Face 提供，并配套发布技术博客。
Z.ai 宣布推出 GLM-5.3，并将其作为开放权重模型提供。用户可以下载模型权重，在本地或其他受支持的环境中运行，并根据需求进行定制。
官方将 GLM-5.3 定位为面向智能体编程和网络防御的高能力模型。相关权重托管在 Hugging Face，Z.ai 同时发布了技术博客，介绍模型及其使用信息。
目前公开材料主要能够确认模型发布和获取渠道；有关具体性能、基准测试及授权条款的细节，仍应以模型页面和官方技术文档为准。
来源核查不代表信息永远正确；请结合原文、观察时间与后续修正判断。
GLM-5.3 is now open-weight. Our most capable model for agentic coding and cyber defense is now available to download, run, and customize.
## John Paul Fusin's post ### John Paul Fusin 6d · GLM-5.3 is now available as an open-weight model for coding, agentic workflows, and cyber defense. Its weights can be downloaded, run, and customized, with local serving paths listed for SGLang, vLLM, Transformers, and other frameworks. Key features: • Post-training focused on complex coding and long-horizon tasks. • Z ai reports a 50% improvement over GLM-5.2 on its in-house Code Bench. • Official benchmark tables cover Terminal Bench 3.0, DeepSWE, CyberGym, and agentic-task evaluations. • The Hugging Face model card lists English and Chinese support, plus Safetensors and FP8 files. May be a graphic of text that says
GLM-5.3 is now open-weight. Our most capable model for agentic coding and cyber defense is now available to download, run, and customize.
Z.ai "GLM-5.3 is now open-weight. Our most capable model for agentic coding and cyber defense is now available to download, run, and customize.
Log inSign up ## Post # Z.ai on X: "GLM-5.3 is now open-weight. Our most capable model for agentic coding and cyber defense is now available to download, run, and customize. Weights: Tech blog: @Zai_org Z.ai @Zai\_org GLM-5.3 is now open-weight. Our most capable model for agentic coding and cyber defense is now available to download, run, and customize. Weights: huggingface.co/zai-org/GLM-5.3 Tech blog: z.ai/blog/glm-5.3 3:04 PM · Aug 28, 20261.4MViews 274 @Zai_org Z.ai @Zai\_org GLM-5.3 is now open-weight. Our most capable model for agentic coding and cyber defense is now available to download, run, and customize. Weights: huggingface.co/zai-org/GLM-5.3 Tech blog: z.ai/blog/glm-5.3 3:04 PM · Aug 28, 20261.4MViews 274 @UnslothAI Unsloth AI @UnslothAI [...] 274 @UnslothAI Unsloth AI @UnslothAI Aug 28 Congrats Z ai team, once again another amazing open-source release! 🥰 We're working on GLM-5.3 Unsloth GGUFs for those who can run it l
# Z.ai releases GLM-5.3 as open-weight model GLM-5.3 is now open-weight ▲ 780jeudesprits268 HN discussion Z.ai releases GLM-5.3 as open-weight model Z.ai has released GLM-5.3, its most capable model for agentic coding and cyber defense, now available as open weights for download, running, and customization. The model is hosted on Hugging Face, with a tech blog detailing its capabilities. The release has sparked community responses, including Unsloth AI working on GGUF versions for local execution and Baseten planning to host it. Notably, GLM-5.3 has moved away from the MIT license to a semi-non-commercial license, similar to other Chinese AI models. [...] > GLM-5.3 is now open-weight. Our most capable model for agentic coding and cyber defense is now available to download, run, and customize. Read original →HN discussionRSS ## HN discussion 268 1. revolvingthrow GLM 5.3 is probably the sweet spot open weights model if you want to go beyond deepseek flash or the new glm flas