An experiment tested six generative large language models against students in an online introductory biomedical and health informatics course. The models scored higher than as many as three quarters of the real-world students in the class.
https://www.sciencedaily.com/releases/2024/09/240916153435.htm