Recent OpenAI agent swarms and large language models have demonstrated rapid progress on deep mathematics after resolving the Navier-Stokes existence and smoothness problem in September 2026, generating hundreds of manuscripts that include verified advances on the quasi-Riemann hypothesis and partial claims toward the Hodge conjecture. These results have tightened trader consensus around the Hodge conjecture as the next full Millennium Prize solution, yet its algebraic geometry complexity, combined with withdrawn papers containing errors and the Clay Institute’s deliberate verification process, keeps the “no solution by December 31, 2027” outcome nearly tied. Birch and Swinnerton-Dyer and Yang-Mills show less reported AI traction, while P versus NP and the full Riemann hypothesis remain distant given their combinatorial and analytic demands. Upcoming model releases and formal Lean checks over the next 14 months represent the key swing factors.