PhysToolBench is a Visual-Language Question Answering (VQA) dataset released in 2025 by the Hong Kong University of Science and Technology (Guangzhou) in collaboration with the Hong Kong University of Science and Technology, Beijing University of Aeronautics and Astronautics, and other institutions. The related research paper is titled "...".PhysToolBench: Benchmarking Physical Tool Understanding for MLLMsThe study aims to evaluate the ability of multimodal large language models (MLLMs) to identify, understand, and create physical tools.

This dataset contains over 1,000 image-text pairs, covering various scenarios including daily life, industry, outdoor activities, and professional environments. It is divided into three levels of difficulty: easy, medium, and hard. The task structure is as follows:

• Tool Creation
• Tool Recognition
• Tool Understanding