DeepSeek delivers frontier-class AI performance at a fraction of competitor prices, but serious privacy concerns make it a ...

适用于常规 AI 模型和 MoE。 DeepSeek 的开源周已经进行到了第三天（前两天报道见文末「相关阅读」）。今天开源的项目名叫 DeepGEMM，是一款支持密集型和专家混合（MoE）GEMM 的 FP8 GEMM 库，为 V3/R1 的训练和推理提供了支持，在 Hopper GPU 上可以达到 1350+ FP8 TFLOPS 的 ...

编程王者Claude地位不稳了？？ 大模型竞技场最新战报出炉，DeepSeek新版R1拿下网页编程第一，小胜Claude Opus 4。 要知道Claude Opus 4可是公认的“全球最强编码模型”。 so，能在编程上战胜Claude Opus 4，DeepSeek-R1-0528到底啥来头？ 看名字你可能以为是个小版本更新，但 ...

Have you ever wished you could have an AI voice assistant that not only understands you but also explains its reasoning, like a thoughtful conversation partner? Whether you’re navigating a busy ...

The Chinese firm has pulled back the curtain to expose how the top labs may be building their next-generation models. Now things get interesting. When the Chinese firm DeepSeek dropped a large ...

最近几天，AI编程工具赛道的战况突然升级，三个消息同时引爆了开发者圈层。 马斯克也下场卷AI编程了 2026年5月初，马斯克旗下xAI正式完成战略整合，更名为SpaceXAI，全力加速进军AI编程工具赛道。很快，一款名为Grok Build的桌面级编程应用意外曝光，定位全桌面 ...

The Chinese start-up used several technological tricks, including a method called “mixture of experts,” to significantly reduce the cost of building the technology. By Cade Metz Reporting from San ...