An AI startup has disclosed that ChatGPT achieved the top scores in this year’s entrance exams for both the University of Tokyo and Kyoto University, surpassing the performance of the highest-scoring human candidates. In the medical program entrance exam under the Natural Sciences III category at the University of Tokyo, ChatGPT scored 50 points higher than the top human score and achieved a perfect score in mathematics. Its performance in the humanities, social sciences, and other science categories also exceeded the highest admission scores. Similarly, at Kyoto University, ChatGPT’s scores in the law school and medical school exams surpassed the highest admission marks. However, ChatGPT scored at a 90% accuracy rate in English, while only achieving a 25% accuracy rate in world history essay questions. In 2024, utilizing the GPT-4 model, ChatGPT failed to meet the admission threshold. Yet, the following year, the o1 model successfully crossed the qualifying line for the first time. The company’s executive believes that enterprises adopting AI should prioritize strategic planning for the future. Experts emphasize that direct comparisons between humans and AI are inappropriate, and that the current format of entrance exams is in urgent need of reform.