GPT-4o

Study Finds AI Models That Account for User Emotions Are More Prone to Factual Errors