Ex-OpenAI researcher's Periodic trains trillion-param Neon on 1300 H200s to beat GPT-6 Astra

LiamFedus · x · 2026-09-16

Periodic, founded by ex-OpenAI researcher Liam Fedus, built high-throughput materials labs in Menlo Park creating a loop where experiments generate fresh data that models learn from and use to decide what to try next. Using only 1,300 H200s plus months of proprietary experimental data, they mid-trained and RL'd an open-source model into "Neon," a trillion-parameter model that beats GPT-6 Astra on their scientific analysis benchmark with relatively little compute. Focus areas include superconductors, magnets, and semiconductor materials.

Related event: Periodic Labs beats GPT-6 on materials science with just 1,300 H200s(41 posts)→

Original post →

More from Companies & People

Companies & People channel →