1,300 H200s and a closed experimental loop: open model Neon beats GPT-6 Astra on materials analysis

OriolVinyalsML · x · 2026-09-16

Liam Fedus' team built high-throughput materials labs in Menlo Park, creating a loop where experiments generate fresh data, models learn from it, and then help decide what to try next. Using only 1,300 H200s plus months of experimental data, they mid-trained and RL'd an open-source model — dubbed Neon — that surpasses GPT-6 Astra on their analysis benchmark. Oriol Vinyals shared the post, calling it welcome progress in AI for science.

Related event: Periodic Labs Trains Science Model Neon on 1,300 H200s, Beating GPT-6 Astra on XRD(59 posts)→

Original post →

More from Research

Research channel →