My Meta friends say it's benchmaxxed af
replyWe used to call this "overfitting," but I suppose everything has to be maxxed now. Fitmaxxed?
replyIt doesn't seem benchmaxxed, ARC AGI 2 score is quite bad (42.5%, GPT 5.4 is 76.1%) and coding is okay. But maybe this is the best Meta can do even benchmaxxing
replyThe impressive part is multimodality, very plausible since there's less focus there by other labs (especially Anthropic)
Given llama 4 mucked up benchmark numbers, I’d take spark announcement with a many grains of salt.
reply