
量子位报道小米 MiMo-V2.6 的强化学习训练过程首次公开直播:Pro 与 Flash 两款模型的成本、步数与评测曲线实时可查,罗福莉说明团队在计算量、环境与评分三个方向上扩展 RL 规模。
转载说明:本文转载自微信公众号「量子位」官方账号(biz_id:MzIzNjc1NzUzMw),原文发表于 2026 年 9 月 17 日 08:56,作者署名「梦晨 发自 凹非寺 量子位 | 公众号 QbitAI」。原文链接:[量子位微信公众号原文](http://mp.weixin.qq.com/s?__biz=MzIzNjc1NzUzMw&mid=2247923967&idx=1&sn=f17eab60e3054b38931194ad3766a2a4),同题文章另见量子位官网 qbitai.com。


Nearly half a year of silence. We spent it studying one problem: how far RL can scale.MiMo-V2.6 is in the middle of its RL run right now. Three things we scaled: compute (~2B tokens per step, 1568 prompts × 16 rollouts, fully async), environments and harnesses (multi-task agentic RL, mixed across multiple harnesses in one run), and grader compute (agentic in-group credit assignment, with test-case and rubric-based rewards). We'll open-source the details piece by piece over the coming weeks.


当预训练的 Scaling 已经发展多年之后,能否把模型与环境交互、产生经验、评价经验再到学习经验的整个 RL 循环,也变成一台可以持续扩大规模的机器?
mp.weixin.qq.com
mimo.xiaomi.com
This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.