Models
Mathematicians dispute OpenAI's Navier-Stokes proof
Independent researchers say OpenAI's 166-page automated solution to the Navier-Stokes problem fails to follow the reasoning methodology claimed in its documentation.
OpenAI submitted a 166-page mathematical paper proposing a solution to the long-standing Navier-Stokes existence and smoothness problem, but outside mathematicians are disputing the validity and clarity of the work. Multiple research teams analyzing the submission report that the mathematical steps in the proof depart significantly from the procedural methods that OpenAI outlined in its technical overview.
The discrepancy has renewed debate over the verification of frontier reasoning models in formal mathematics. Because neural networks generate proofs without transparent mechanistic explanations, researchers reviewing the 166-page text have struggled to reconstruct the origin of key conceptual leaps or verify whether intermediate lemmas hold under standard analytical scrutiny.
OpenAI has not publicly clarified the differences identified by independent reviewers. The Navier-Stokes equations, one of the Clay Mathematics Institute Millennium Prize Problems, have resisted rigorous solution for decades, and mathematicians warn that verifying machine-generated candidates requires exhaustive formal decomposition before any claims of resolution can be accepted.