Z.ai 发布了 GLM-5.2 开源模型,具备 1M 上下文窗口和两种推理模式,专为智能体和长周期编码任务设计
多源视角
RT Kilo<br>GLM-5.2 runs in Kilo Code on day one. The 1M context window and Max effort mode are both live. Point your config at it and go!<hr style="border:0;border-top:1px solid #8…
查看这条来源RT ollama<br>🤯 GLM-5.2 is here — built for long-horizon coding and agentic tasks, now with a solid 1M-token context. <br><br>The strongest open-source coding model yet!<br><br>Ava…
查看这条来源RT friendliai<br>GLM-5.2 is live on Friendli Model API as a Day 0 partner.<br><br>http://Z.ai's newest flagship for agentic engineering brings:<br>🔹 Significant improvements in co…
查看这条来源RT LMSYS Org<br>🎉 Meet GLM-5.2 from @Zai_org, the new flagship for long-horizon tasks built on a solid 1M-token context. Day-0 support is now live in SGLang!<br><br>✅ Advanced cod…
查看这条来源RT Phala<br>Congrats to @Zai_org on GLM-5.2 — open weights, 1M context, and stronger coding/agentic performance are a big step for open AI.<br><br>Phala is proud to be a launch par…
查看这条来源RT Venice<br>GLM 5.2 is now live on Venice.<br><br>@Zai_org's frontier open-source model built to power agents and long-horizon agentic coding, with High and Max reasoning modes fo…
查看这条来源AI 摘要
Z.ai 发布了 GLM-5.2 开源模型,具备 1M 上下文窗口和两种推理模式(High 和 Max),专为智能体和长周期编码任务设计。该模型在编码和智能体任务上有显著改进,已上线 Venice 平台,对 Pro 用户完全私有可用。 核心观点: 1. GLM-5.2 提供两种推理模式:GLM-5.2 (max) 和 GLM-5.2 (high),分别对应极限推理和高效推理。 2. 模型支持 1M 上下文窗口,专为长周期智能体编码任务设计。
推荐理由高信息密度,值得细读
原文
无法获取正文,可打开下方原始链接查看。
讨论
暂无评论。