MM-IQ: Benchmarking Human-Like Abstraction and Reasoning in Multimodal Models Paper • 2502.00698 • Published 3 days ago • 18