WHAT NEW IN X11.05 THU

AI collections from X · 22 collections · 67,905 posts

gpt-6POST

Football analysis model test: GPT-6 Astra scores 0% success

The author says GPT-6 Astra scored 0% on the same football analysis-modeling tasks measured under real conditions on their open-source eval bench.

Max
@MaxSebtiArchived 2026.10.05Originally posted 2026.10.05
Max @MaxSebtiOriginally posted 2026.10.05
Open original on X

we measured the same success rate on football (soccer) analysis-model work GPT-6 Astra scored 0% when measured under real conditions on our open-source eval bench

This post quotes another post: “Livestream: GPT-6 Astra plays StarCraft against top bot Pluto”

x.com/maxsebti/status/2106820324536803492Archived 2026.10.05 10:05