EmoBench-M is a benchmark dataset proposed by Shenzhen University, Guangming Laboratory, University of Macau and other institutions in 2025 to evaluate the sentiment understanding ability of multimodal large language models (MLLMs). The related paper results are "EmoBench-M: Benchmarking Emotional Intelligence for Multimodal Large Language Models", which aims to fill the gaps in existing unimodal or static emotion datasets in dynamic and multimodal interaction scenarios, and to be closer to the complexity of human emotional expression and perception in real environments.

This dataset contains video clips of people expressing different emotions, covering 13 real-world scenarios. It aims to comprehensively assess emotional intelligence capabilities across three key dimensions: basic emotion recognition, conversational emotion understanding, and analysis of complex social emotions. Each video sample is accompanied by corresponding dialogue prompts to guide the model in identifying and judging the emotions expressed.