OpenAI says Navier-Stokes proof used a next-gen model far stronger than GPT-6 Astra

According to Axios, the model OpenAI used to claim a solution to the Navier–Stokes Millennium Prize Problem was an internal model "significantly stronger than GPT-6 Astra," not the publicly released Astra. Omar Sanseviero of dair-ai relayed the official statement that the proof was produced by a set of agents using OpenAI's next-generation model. This means an unreleased next-gen model sits behind the Millennium Problem claim, and the Astra the public sees is not OpenAI's capability ceiling.

Confirmed

Why It Matters

2026-09-09 ~ 2026-09-09 · 6 related posts

Full story(4 episodes)→

Primary sources