EmoBench-M introduces the first comprehensive benchmark for evaluating emotional intelligence in multimodal large language models, enabling more human-like AI interactions.

• Addresses the gap in measuring dynamic, multimodal emotional understanding
• Enables robots and AI systems to better recognize and respond to human emotions
• Creates a foundation for embedding emotional awareness in next-generation AI applications
• Provides metrics for comparing emotional intelligence across different MLLMs

Medical Impact: This research has significant implications for healthcare, where AI systems with emotional intelligence could improve patient experiences, enhance mental health applications, and enable more empathetic telehealth interactions.

EmoBench-M: Benchmarking Emotional Intelligence for Multimodal Large Language Models