Recent breakthroughs have sharply accelerated AI progress on open mathematical problems. In September 2026, OpenAI used autonomous AI agents powered by an internal frontier model to produce a Lean-verified solution to the Navier-Stokes existence and smoothness problem, one of the Millennium Prize challenges. Early October saw the company release 722 papers detailing full or partial solutions to over 370 longstanding conjectures across number theory, algebra, topology, and optimization, including multiple Erdős problems and advances on other Millennium questions. Smaller labs such as doubleAI have also delivered Lean-formalized improvements on decades-old bounds. These results, achieved through iterative large language model generation paired with formal verification, have shifted trader focus toward which additional high-profile problems frontier systems may resolve before year-end.