非常厉害!
从整个训练过程来看,估计每百万 token 训练成本:
Pro:约 34.3 美元/百万 token Flash:约 8.7 美元/百万 token
当前训练总成本已经超过 110 万美元,所以 MiMo-V2.6 的 Pro + Flash 训练总成本在 300 万美元?
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 scale...