On April 27, 2026, LifePrompt, a Japanese AI firm, unveiled exam results demonstrating that ChatGPT, powered by OpenAI's latest model, outperformed the highest-scoring human candidates across several subjects in the undergraduate entrance exams for both the University of Tokyo and Kyoto University. In the science section of the University of Tokyo's exam, ChatGPT scored 503 out of 550, surpassing the admission threshold by a significant margin of 50 points. At Kyoto University's School of Medicine, it achieved an impressive 1176 points, outperforming the top human scorer by 78 points. The exams incorporated image-based question inputs and were graded manually by educators from Kawai Juku. Notably, in 2024, this same AI model fell short of passing all entrance exams at the University of Tokyo, marking a remarkable turnaround with its recent achievements.
