关于这个文章区
这个文章区是我的AI 开发者工具箱的延伸：工具给数字，文章给判断。
内容分三类：评测榜盘点（LLM 评测榜、CPU 跑分天梯、显卡 AI 算力排行）、硬件实测（在核显、NPU、纯 CPU 上跑大模型的速度记录与避坑）、技术科普（Token 是什么、流式输出原理、MoE 架构等基础概念）。
所有内容均为我自己的第一手整理与实测，数据标注来源，观点仅供参考。如果你在本站工具里查到了数字，想知道"这意味着什么"，来这里看文章最合适。
About this blog
This blog is the extension of my AI Developer Toolbox: the tools give numbers, the articles give judgment.
Three categories: benchmark roundups (LLM leaderboards, CPU ladders, GPU AI compute rankings), hands-on hardware tests (running LLMs on iGPUs, NPUs and CPU-only machines — speeds and pitfalls), and explainers (what a token is, how streaming works, MoE architecture).
All content is first-hand — measured or curated by me, with sources noted. If a number in one of our tools made you wonder "so what does it mean", this is the place to read.