ChatGPT Surpasses Top Human Scorers in Tokyo and Kyoto University Entrance Exams
Using the latest 'ChatGPT-5.2 Thinking' model, AI achieved 'Top of Class' scores in entrance exams for Japan's elite universities, marking a leap in generative AI capabilities.
📋 Article Processing Timeline
- 📰 Published: April 28, 2026 at 14:42
- 🔍 Collected: April 28, 2026 at 15:02 (19 min after Published)
- 🤖 AI Analyzed: April 28, 2026 at 15:31 (29 min after Collected)
The latest generative AI, 'ChatGPT-5.2 Thinking' developed by OpenAI, has achieved scores exceeding the top human applicants in this year's entrance exams for the University of Tokyo and Kyoto University. According to Tokyo-based startup Life Prompt, the AI scored 503 out of 550 points in the science category at Todai, surpassing the human top score of 453. Most notably, the AI scored perfectly in the extremely difficult mathematics exam. While it excelled in objective and logical subjects, it struggled with subjective essay questions in world history, scoring only 15 out of 60. This leap from failing all subjects in 2024 to becoming a 'top scorer' in just two years highlights the rapid evolution of generative AI logic and image-to-text processing.