Nathan Lambert· @natolambert · X·· 2 小时前同事件AI 评分67
AI 导读
RicursiveAI 发布从零预训练的 65B 稠密大语言模型 Ripple-65B Preview，称其在智能体编码任务上超过 Muse Spark Open-816B、Beam-501B、Inkling-975B 和 Nemotron 3 Ultra-550B 等更大模型。Nathan Lambert 转发时表示欣赏这个精简项目，希望更多美国实验室效仿，并指出它在思路上延续了 olmo 系列。
同一件事，精选收录的是《RicursiveAI 发布 Ripple-65B Preview，训练成本 3200 万美元》
正文 · 原文
Love this scrappy project. Hoping more US labs follow suit, showing they can also make solid models with small focused teams. This also intellectually follows on the olmo lineage, so super cool to see what our baby looks like scaled up.
Today we release Ripple-65B Preview, our 65B dense LLM we pretrained from scratch. Ripple sets a new state-of-the-art for U.S. open-weight models in agentic coding tasks, surpassing significantly larger models including Muse Spark Open-816B, Beam-501B, Inkling-975B, and Nemotron 3 Ultra-550B. Learn more: https://www.ricursive.com/ripple在 X 查看被引用的帖子
来源：Nathan Lambert · x.com