Fetching the paper…

MM-IQ: Benchmarking Human-Like Abstraction and Reasoning in Multimodal Models · Around