Moonshot AI· KimiResearch paper

WorldVQA

Moonshot AI releases WorldVQA, a benchmark of 3,500 image-question pairs to evaluate factual visual world knowledge in multimodal LLMs, with a focus on head vs. tail distribution. Experiments show frontier models often fall below 50% accuracy, and the dataset and evaluation scripts are open-sourced.

From the source

We are releasing WorldVQA , a new benchmark designed to measure the factual correctness of Multimodal Large Language Models (MLLMs).

kimi.ai

Announced

03 Feb 2026 00:00 UTC

Detected

28 Aug 2026 13:24 UTC (+206d after)