Press Space to continue
Frontier AI Models Jump to 30 Percent on Abstract Reasoning Test
Press Space to continue
ARC-AGI-3 was built to be brutal. Humans solve 100% of the environments. As of March, every frontier model was under 1%. Opus 5 just put up 30.2%. Not quite solved, but the jump is very real!
open on x · 1 viewViews and clicks count once per visitor per day. Clicks are opens of this post on X.