OpenAI's new model reportedly went from RL restart to Navier–Stokes solution in two weeks

haider1 · x · 2026-09-10

haider1 lays out a striking timeline for OpenAI's new model: RL training reportedly restarted on Aug 28, the model was set to work on Millennium Prize Problems by Sep 1, and a new checkpoint after Sep 3 produced a Navier–Stokes solution — a leap in capability within roughly two weeks of training.

Quoted posts call the model a "newborn" that is already better at math than humans. Note: third-party account, not officially confirmed by OpenAI.

Related event: OpenAI's Navier-Stokes Model Trained Only Days Before 88-Hour Run(2 posts)→

Original post →

More from Models

Models channel →