Attached: 1 image
from here. seems like GPT-4o has the same problems as all the others. some nice benchmarks, but people who are actually using it say it's pretty mediocre. https://old.reddit.com/r/LocalLLaMA/comments/1crkaju/anyone_else_slightly_underwhelmed_with_gpt4o/