WHAT NEW IN X10.31 SAT

AI collections from X · 22 collections · 63,170 posts

gpt-6IMAGE

GPT-6 Sol and Luna benchmark charts called the hardest to read yet

A user shares screenshots of the GPT-6 Sol and Luna benchmarks and calls them the worst, most unreadable benchmark charts in OpenAI history.

alex getman
@alexgetmancomArchived 2026.09.23Originally posted 2026.09.23
GPT-6 Sol and Luna benchmark charts called the hardest to read yet (image 1 of 4)
GPT-6 Sol and Luna benchmark charts called the hardest to read yet (image 2 of 4)
GPT-6 Sol and Luna benchmark charts called the hardest to read yet (image 3 of 4)
GPT-6 Sol and Luna benchmark charts called the hardest to read yet (image 4 of 4)
alex getman @alexgetmancomOriginally posted 2026.09.23
Open original on X

GPT-6 Sol and Luna benchmarks are out. These are the worst benchmarks in OpenAI history. They’re basically impossible to read. If anyone can make sense of them, let me know.

x.com/alexgetmancom/status/2102463388895035704Archived 2026.09.23 10:17