The Openai experimental model has produced gold at the Olympiad of international mathematics

Openai has achieved “golden performance” at international Olympiad in mathematics, marking another important step for the rapid growth of AI. Alexander Wei, a researcher at Openai working on LLM and reasoning, published on X that an experimental research model has delivered this “great long -time challenge in AI”.

According to Wei, an unprecedented model of Openai was able to solve five problems out of six on one of the oldest and most prestigious mathematics in the world, earning 35 out of 42 points in total. The International Mathematics Olympiad (IMO) sees countries sending up to six students to solve extremely difficult algebra and pre-capping problems. These exercises are apparently simple but generally require a certain creativity to mark the highest marks on each problem. For this year’s competition, only 67 of the 630 candidates in total received gold medals, or about 10%.

AI is often responsible for tackling complex data and repetitive actions, but it is generally lower than problem solving that require more creativity or complex decision -making. However, with the latest OMI competition, Openai says that his model has managed complex mathematical problems with human reasoning.

“In doing so, we have obtained a model that can develop complex and waterproof arguments in terms of human mathematicians,” wrote Wei on X. Wei and Sam Altman, CEO of Openai, both added that society does not expect to free anything with this level of mathematical capacity for several months. This means that the next GPT-5 will probably be an improvement compared to its predecessor, but it will not have the same impressive ability to compete in the OMI.