ChatGPT-4o, ChatGPT-4 and Google Gemini are compared with Students: A Study in Higher Education

Abstract

The rapid advancement of artificial intelligence technologies, their pervasive use in every field, and the growing understanding of the benefits they bring have led actors in the education sector to pursue research in this field. In particular, the use of artificial intelligence tools has become more prevalent in the education sector due to the increasing number and functionality of these tools. The educational research conducted in the 4o version, which is relatively new in the context of ChatGPT tools that have gained popularity, has not yet produced a substantial body of evidence. In this study, ChatGPT-4, 4o and Google Gemini were subjected to a physiology examination taken by veterinary department students. A comparative analysis of the students' performances and the performances of these tools was conducted. In the comparison within the artificial intelligence tools themselves, the latest ChatGPT-4o version showed an accuracy rate of 90%. It was followed by Gemini and ChatGPT-4. All of the tools used in the physiology course exam were able to pass the exam with a satisfactory level. ChatGPT-4o, 4 and Gemini outperformed the students.