TOKYO 29 April 2026 (The Capital Post) – OpenAI’s ChatGPT has reportedly achieved scores higher than the best-performing human candidates in the entrance examinations for two of Japan’s most prestigious universities, marking a major milestone in the advancement of artificial intelligence technology.
According to reports, the AI model was tested on the 2026 entrance examinations for the University of Tokyo and Kyoto University by Japanese AI venture firm LifePrompt Inc. The chatbot used was OpenAI’s ChatGPT 5.2 Thinking model, which answered questions without internet access and relied entirely on reasoning and previously learned knowledge.
The results showed that ChatGPT exceeded the highest scores achieved by successful applicants in several competitive programmes, including the University of Tokyo’s highly challenging Natural Sciences III medical track examination. The AI also reportedly obtained perfect marks in mathematics during the assessment.
The examinations included essay-based questions, which were graded by instructors from a major Japanese preparatory school to ensure consistency with real university admission standards. While the AI demonstrated exceptional performance in mathematics and analytical subjects, it showed weaker results in essay-style areas such as world history and written interpretation.
-Advertisement-
The achievement represents a dramatic improvement compared to earlier testing in 2024, when previous ChatGPT models failed to meet the minimum passing threshold for some university entrance exams. Researchers noted that the rapid advancement highlights the growing capability of AI systems in complex academic problem-solving.
Experts in artificial intelligence said the results may reignite debate over the future of education and university admission systems, particularly whether traditional exams focused heavily on memorisation and calculation remain suitable in an era of increasingly sophisticated AI technologies. -The Capital Post
