Search Results: AIMathematics

OpenAI announced its experimental language model achieved gold medal-level results on the prestigious International Mathematical Olympiad using plain-text reasoning, matching a standard few humans reach. The claim, self-graded and released early against IMO embargo requests, raises questions about verification and scientific protocol while signaling a potential leap in AI reasoning capabilities.

Mathematicians reveal surprising insights from closed-door AI testing sessions, exposing both the impressive capabilities and fundamental limitations of reasoning models like ChatGPT and Gemini when tackling advanced math problems. Behind the hype, researchers discovered AI struggles with geometric intuition and genuine problem-solving despite excelling at pattern matching.