MM-IQ: Benchmarking Human-Like Abstraction and Reasoning in Multimodal Models Paper • 2502.00698 • Published 26 days ago • 23 • 2