OpenAI claims Navier-Stokes Millennium Prize proof by agent swarm running a model beyond GPT-6 Astra
danintheory · x · 2026-09-09
OpenAI announced that a group of agents running an unreleased next-generation model—significantly more capable than GPT-6 Astra—produced a proof of the Navier-Stokes Millennium Prize Problem, which asks whether smooth 3D fluid flow can break down and has stood unresolved for roughly 90 years.
ARC Prize's Polynoamial framed the cost trajectory: o3 needed $500,000 to score 87.5% on ARC-AGI 1 at launch, while Astra now scores higher for $20; after OpenAI and Google DeepMind burned enormous compute for IMO gold in 2025, a $20/month ChatGPT subscription could plausibly do the 2026 IMO.
His take: massively scaling test-time compute offers a glimpse of a future where, within a year, everyone has an AI capable of solving problems of this caliber.
More from Models
- OpenAI claims huge Navier-Stokes math discovery, academics cry foul — wiredmagazine · 2026-09-09
- OpenAI used a model 'significantly more capable' than Astra for Navier-Stokes, per Axios — 141_1337 · 2026-09-09
- Dwarkesh on MagicAILabs' 50x compute claim: RSI may be less compute-bottlenecked than we think — AccBalanced · 2026-09-09
- Rumor: OpenAI may have cracked Navier-Stokes, with human mathematicians laying years of groundwork — LucaAmb · 2026-09-09
- '50% of open problems just solved' — commentator marvels at frontier model progress — BorisMPower · 2026-09-09
- Cheap models via OpenRouter fall apart in agentic harnesses: GLM and DeepSeek can't match Claude — scottyLogJobs · 2026-09-09