MM-IQ: Benchmarking Human-Like Abstraction and Reasoning in Multimodal Models Paper • 2502.00698 • Published 11 days ago • 22 • 2