Open has achieved “gold medal level performance” at the International Mathematics Olympiad, which has exhibited another important milestone for AI’s rapid development. Alexander V, LLMS and reasoning, a research scientist at Openi, posted on x That an experimental research model presented this “Grand Grand Challenge in AI”.
According to V, a non -cheerful model of Open has been able to solve five of the world’s longest and five of the most prestigious mathematics competitions, scoring 35 points out of a total of 42 points. The International Mathematics Olympiad (IMO) saw that the countries send six students to solve the problems of the most difficult algebra and calculus. These exercises are seemingly simple, but usually require some creativity to get the most numbers on each issue. For This year’s competitionOnly 67 out of the total 630 contestants earned gold medals, or about 10 %.
AI is often assigned to deal with complex datases and repeated measures, but it is usually short -lived when it is commonly talked about solving issues that require more creativity or complex decision -making. However, with the latest IMO competition, Open says its model has been able to handle complicated mathematics issues with human reasoning.
“By doing so, we have achieved a model that can produce complicated, water tight arguments at the level of human mathematics,” said the company does not expect to release anything with this level of mathematics for several months. This means that the incoming GPT5 will likely improve its predecessor, but it will not feature such an impressive ability to compete in the IMO.