OpenAI’s early October 2026 release of 722 papers claiming solutions or advances on 377 open problems, including progress toward the Riemann hypothesis and expansions on a recent Navier-Stokes result, has sharply accelerated expectations for AI-driven mathematical discovery in 2026. The internal frontier model produced many Lean-formalized proofs, with some verified independently, though mathematicians note variable quality, occasional retractions, and heavy reliance on existing human work. Competing efforts, such as doubleAI’s low-cost improvements on longstanding graph theory bounds, highlight intensifying lab competition and the growing role of formal verification tools. Upcoming scrutiny of these results and further model releases could quickly shift which specific problems traders view as likely to reach full, accepted resolution before year-end.