
自媒体视频动效 · 人物后大字接全屏宣言
@loktar00模型对比
Kimi K3.1 Pro at High vs Deepseek Flash at Max Kimi looks good, but Deepseek actually in my opinion did it better at a much lower parameter count and cost. I did this via Kimis web harness I have allegro, it used 1.3% of my usage. Deepseek was DSH.
模型对比新0:37
模型对比新0:20
模型对比新0:28
模型对比新0:11
模型对比新0:10
模型对比新0:22