精选理由
小红书开源自研大模型，支持多模态理解，与华为Ascend深度适配，为内容平台构建AI基础，值得关注。
小红书发布自研大模型dots3-note，MoE模型，总参数280B，激活参数16B，512K上下文，支持文本、视觉和语音理解。以Apache 2.0协议在Hugging Face和GitHub上开源，华为Ascend零延迟适配，标志着月活跃用户超3亿的平台上，AI从功能转变为基础。
原文 · Pandaily
Xiaohongshu Quietly Open-Sources Its Own Large Model: Content Platforms Build Their Own Base
Xiaohongshu released a large model overseas, the dots3-note preview, a MoE model with 280B total and 16B activated parameters, 512K context, and text, vision, and voice understanding. Released under Apache 2.0 on Hugging Face and GitHub with Huawei Ascend adapting on day zero, it signals that the platform with over 300 million monthly users is turning AI from a feature into a foundation.