Mozi (@yeahhe) 在 实测Poe的 DeepSeek-R1 模型能力最接近官方网页端,上下文长,Max token 33k,数学难题思考不会中断 中发帖
[image]
[image]
Poe的DeepSeek-R1由together ai提供
硅基流动、英伟达云、together ai的免费版本都是思考过程就中断
[image]
[image]
Poe的DeepSeek-R1由together ai提供
硅基流动、英伟达云、together ai的免费版本都是思考过程就中断