WHAT NEW IN X11.06 FRI

AI collections from X · 23 collections · 69,408 posts

gpt-6IMAGE

JevBench ranks Kimi K3 first and GPT-6 Astra second to last

A benchmark built from Jev's opinions on each model puts Kimi K3 on top and GPT-6 Astra second to worst, with OpenAI faring poorly.

Adam Holter
@AdamHolterererArchived 2026.09.23Originally posted 2026.09.22
JevBench ranks Kimi K3 first and GPT-6 Astra second to last
Adam Holter @AdamHolterererOriginally posted 2026.09.22
Open original on X

Introducing JevBench: What does Jev think about each model? Jev thinks Kimi K3 is the best model, and GPT-6 Astra is the second to worst. OpenAI takes a whooping on JevBench

x.com/adamholtererer/status/2102392507325927846Archived 2026.09.23 10:29