OpenAI’s September 8, 2026 announcement of a finite-time singularity for the forced 3D Navier-Stokes equations, backed by an analytical write-up and Lean formalization produced by roughly 10,000 internal agents on an unreleased model, underpins the 94% market-implied probability that the company will not retract the result this year. The Lean verification and public repository provide machine-checked rigor that has withstood initial scrutiny, while OpenAI’s prompt correction of unrelated sign errors in its October 6 batch of preprints demonstrates a pattern of addressing flaws without abandoning core claims. Lingering priority disputes and a noted mismatch between the natural-language proof and Lean code introduce uncertainty, yet these have not invalidated the central construction. A late-year discovery of a substantive gap that survives community review could still trigger retraction, though current evidence favors stability through year-end.