OpenAI’s October 2026 release of 722 AI-generated mathematical papers, including verified solutions to hundreds of open research problems and progress toward Millennium Prize challenges like the quasi-Riemann hypothesis, has sharply elevated expectations for additional Epoch AI-classified FrontierMath Open Problems resolutions before year-end. Leading models such as GPT-5.6 variants now achieve 80-90% on FrontierMath tiers, building on earlier verified solutions like the Hadamard matrix and hypergraph problems. Competitive pressure from Anthropic’s Claude and Google DeepMind systems, combined with multi-agent scaffolding and Lean formalization, continues to accelerate discoveries. Traders are monitoring Epoch AI’s ongoing classifications and any further official announcements through December, as these directly determine market resolution.