gemini-4POST
DeepSWE 榜首换人就被说 benchmark 饱和
网友吐槽双标：Gemini 3.8 Flash 登顶 DeepSWE 时被说榜单饱和，Sol 登顶就是 SOTA，Gemini 4 Argon 登顶又变成榜单饱和。
Harshith
@HarshithLucky328 天前收录原推发布于 2026.10.01
Harshith @HarshithLucky3原推发布于 2026.10.01
that guy: when Gemini 3.8 flash topped DeepSWE - benchmark is saturated then 6.1 sol topped it - SOTA in DeepSWE Gemini 4 Argon topped it - ohh no benchmark is saturated this is not just about DeepSWE
x.com/harshithlucky3/status/2105424982071541888采集于 2026.10.01 09:59