Periodic Labs trains Neon with just 1,300 H200s, beating GPT-6 Astra on materials analysis

LiamFedus · x · 2026-09-16

Periodic Labs has built high-throughput autonomous materials labs in Menlo Park that close the loop between experiments and models: the labs generate fresh data, models learn from it, and then help decide what to try next.

Using only 1,300 H200 GPUs plus months of proprietary experimental data, the team mid-trained and RL'd an open-source model into Neon, which surpasses GPT-6 Astra on their analysis benchmark. The footage is real lab video.

Initial focus is on hard materials-science problems: superconductors, magnets, and semiconductor materials.

Related event: Periodic Labs Unveils Neon: 1,300 H200s Train a Science Model Beating GPT-6(13 posts)→

Original post →

More from AGI Musings

AGI Musings channel →