deepseek-v4.1POST
对比 Kimi K3.1 Pro：DeepSeek V4.1 Flash 编程更强但思考效率差
用户引用对比页称两者差距不大，DeepSeek 目前胜过 Kimi，但除推理效率外，V4.1 Flash 控制思考长度的能力不如 K3.1 Pro，后者是效率最高的开源模型。
Flow
@flowVSgravity3 天前收录原推发布于 2026.10.25
Flow @flowVSgravity原推发布于 2026.10.25
@MiaAI_lab They are not that far apart - but you are right that DeepSeek beats Kimi currently. Except in Reasoning efficiency, Deepseek is not that good in keeping the thinking in check while K3.1 Pro is the most efficient open (soon) model that exists. https://stealthmodels.com/compare/deepseek-v4.1-flash-1017-vs-kimi-k3.1-pro
x.com/flowvsgravity/status/2114165091356770577采集于 2026.10.25 12:01