WorldVQA

We introduce WorldVQA, a benchmark designed to evaluate the factual correctness and atomic vision-centric world knowledge of Multimodal Large Language Models (MLLMs). Current evaluations often conflate visual knowledge retrieval with reasoning. In contrast, WorldVQA decouples these capabilities to strictly measure "what the model memorizes." The benchmark assesses the atomic capability of grounding and naming visual entities across a stratified taxonomy, spanning from common head-class objects to long-tail rarities. We hope WorldVQA serves as a rigorous test for visual factuality, thereby establishing a standard for assessing the encyclopedic breadth and hallucination rates of current and next-generation frontier models.