OpenAI’s early October release of 722 AI-generated math papers from an unreleased internal frontier model, following its September claim of a Lean-verified solution to the Navier-Stokes Millennium Prize Problem via 10,000 coordinated agents, has driven strong trader focus on rapid AI progress. The papers advance or resolve hundreds of open problems across number theory, algebra, topology, and graph theory, including partial results tied to the Riemann hypothesis. Rivals such as Anthropic and smaller labs like doubleAI have similarly formalized advances on Erdős problems and other conjectures using multi-agent systems and Lean verification. Mathematicians continue reviewing the scale of outputs amid debates over verification quality, attribution, and whether these mark sustained capability gains or isolated compute-heavy efforts ahead of year-end benchmarks.