insane. Competing with Qwen 3 (latest) while using ~6T tokens, and all in the open. They need to refine OLMoE recipe now, and they may well become a mainstream choice for research.
正在載入線程內容
正在從 X 取得原始推文,整理成清爽的閱讀畫面。
通常只需幾秒鐘,請稍候。
正在載入線程內容
正在從 X 取得原始推文,整理成清爽的閱讀畫面。
通常只需幾秒鐘,請稍候。
共 1 則推文 · 2025年11月20日 下午3:45
insane. Competing with Qwen 3 (latest) while using ~6T tokens, and all in the open. They need to refine OLMoE recipe now, and they may well become a mainstream choice for research.