OpenAI internal math model
by OpenAI · Model · Language & code models
Internal OpenAI model, 10,000 agents for 88 hours, found finite-time blowup for forced viscous Navier-Stokes case
StatusAnnouncedAccessClosed
Evidence
- Oct 11Unreleased internal model said to be far stronger than GPT-6 Astra produced the manuscripts
- Sep 8Behind the Navier–Stokes result: about 10,000 concurrent agents, 4.9 million messages, roughly 300 billion output tokens, over about 5 days
- Sep 8Commenters estimate about $15M at API rates for output tokens alone; the model is internal-only, so real cost is unknown