2 stories in this blend

Discussion emerged around the training techniques used for the system that solved the Navier-Stokes math problem. OpenAI clarified that performance improvements came from extensive reinforcement learning on an existing foundation model rather than a completely new pretraining run.

Researchers at OpenAI claim an unreleased model successfully solved the Navier-Stokes existence problem, one of seven Millennium Prize Problems. Working across 88 hours with roughly 10,000 agents, the system generated a mathematical proof that was later checked using formal verification software.