WHAT NEW IN X10.28 WED

AI collections from X · 22 collections · 58,924 posts

minimax-m3.1VIDEO

MiniMax M3.1 vs DeepSeek V4.1 Flash and two rivals on the same volcano prompt

A four-model volcano prompt test ranks Fable 5.1 (medium) first, MiniMax M3.1 (max) second, DeepSeek V4.1 Flash third for poor lighting, and GPT 5.6 Sol last for degraded quality.

FeiZ
@Fei2411Archived 7 days agoOriginally posted 2026.10.20
FeiZ @Fei2411Originally posted 2026.10.20
Open original on X

MiniMax M3.1 的实测来了,拿 MiniMax M3.1(max) 和 DeepSeek V4.1 Flash(high)、GPT 5.6 Sol(xhigh)、Fable 5.1(medium)做了个火山对比。 综合来看 M3.1 表现不错的。5.6 Sol 已经降智到没法用了....V4.1 Flash 好像遇到了点问题,其实能看出来它喷发做的还行,但是最后光线没调好,实在有点看不清;Fable 5.1 无需多言,细节依然碾压,而且只开了 medium.... 综合来看 Fable 5.1>MiniMax M3.1>DeepSeek V4.1 Flash>GPT 5.6 Sol 我还做了些测试,明天给你们分享更多!

x.com/fei2411/status/2112572281688060374Archived 2026.10.21 12:05