AI Surpasses Human Top Performers: ChatGPT Achieves Highest Scores in Japanese University Entrance Exams
Today, Japanese AI company LifePrompt announced test results showing that ChatGPT, based on OpenAI's latest model, scored higher than human top performers in multiple subjects on undergraduate entrance exams for the University of Tokyo and Kyoto University.
It is reported that ChatGPT scored 50 points higher than the highest-scoring human candidate in the highly competitive natural science category (medical major) entrance exam for the University of Tokyo, and achieved a perfect score in mathematics.
In 2024, this AI was unable to pass all of the university’s entrance exams, but has now achieved a breakthrough.
ChatGPT also participated in this year’s national unified university entrance exam in Japan, and LifePrompt compiled and statistically analyzed its total scores.
For the University of Tokyo’s humanities and social sciences exam, which has a total score of 550, ChatGPT scored 452; for the science exam, also with a total score of 550, it scored 503.
Both scores are higher than the highest scores of admitted students announced by the university: the highest score in the humanities and social sciences category was 434, and the highest score in the science category was 453.
In the Kyoto University exams, the AI scored 771 on the law school entrance exam, exceeding the highest admitted score of 734; and scored 1176 on the medical school exam, higher than the human top performer’s score of 1098.