An experimental model from OpenAI won a gold award at the International Mathematical Olympiad, correctly solving 5 out of 6 problems and scoring 35 out of a possible 42 points. Such a result demonstrates the level of the strongest mathematical students from around the world and indicates significant progress in the development of logical thinking of artificial intelligence. About this informs Engadget.
The model, which has not yet been publicly presented, underwent a simulation of participation in the Olympics without human intervention. She not only solved problems correctly, but also formulated complete mathematical proofs. OpenAI described it as a “long-awaited breakthrough in reasoning” that the team has been working toward for several years.
The International Mathematical Olympiad is one of the most prestigious competitions for schoolchildren, which brings together more than 600 participants from dozens of countries every year. In 2025, only 67 of them won gold, which is about 10% of all participants. Artificial intelligence from OpenAI demonstrated the result exactly at the level of this category. Despite this, the company currently does not intend to release the model into open access. According to Wei, it will not yet be integrated into consumer products.
OpenAI has been working for a long time to create models capable of abstract thinking and logical analysis. In this context, the International Mathematical Olympiad was considered one of the highest standards for testing such possibilities. The obtained result confirms that AI is gradually mastering areas that until recently were considered an exclusively human prerogative.
