Science ❯Mathematics

Problem Solving

Competitions Data Contamination Sofa Problem Proof Techniques Matrix Multiplication Combinatorics AI Performance Logic

Google and OpenAI Claim Unofficial Gold at Math Olympiad, Prompting Benchmarking Debate

IMO organizers stress that AI’s gold-level performance is unofficial, with no formal endorsement or benchmark status.

DeepMind Unveils AlphaEvolve, an AI That Redefines Algorithm Discovery

Microsoft Unveils Phi-4, a Small AI Model Excelling in Complex Reasoning

Mathematician Claims Breakthrough in Decades-Old Moving Sofa Problem

FrontierMath Benchmark Reveals AI's Struggles with Complex Math