Openai’s experimental model won gold at the International Mathematics Olympics
Openai achieved a “gold medal-level performance” at the International Mathematics Olympics, laying the foundation for another important milestone in the fast-paced growth of AI. OpenAI research scientist Alexander Wei posted LLM and reasoning on X on X, who posted an experimental research model on X that provides an experimental research model with the “Long-term Challenge in AI.”
According to Wei, an unreleased model from Openai was able to solve five of six problems in one of the world’s longest and prestigious math competitions, earning a total of 35 of 42 points. The International Mathematics Olympic Games (IMO) believes that countries send up to six students to solve extremely difficult algebra and pre-line problems. These exercises may seem simple, but usually require some creativity to score the highest on each question. In this year’s competition, only 67 of the 630 players won gold medals, about 10%.
AI is often responsible for dealing with complex data sets and repetitive operations, but is often lacking when solving problems that require more creativity or complex decisions. However, in the latest IMO competition, Openai said its model is able to handle complex mathematical problems through human-like reasoning.
“In doing so, we get a model that can formulate complex watertight arguments at the level of human mathematicians,” Wei wrote on X. Wei and OpenAI CEO X. “This means that the upcoming GPT-5 may be improved from its predecessors, but it won’t have the same impressive capabilities that compete in IMO.
