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

LiamFedus · x · 2026-09-16

Periodic Labs built high-throughput materials labs in Menlo Park to close the loop between experiments and models: labs generate fresh data, models learn from it and then suggest 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 called Neon that surpasses GPT-6 Astra on their analysis benchmark. Initial focus: superconductors, magnets, and semiconductor materials. Dwarkesh, after visiting the lab, noted how wide the materials-synthesis search space is yet how amenable it is to depth-first search, since each experiment's design improves as data piles up.

Related event: Periodic Labs beats GPT-6 on materials science with just 1,300 H200s(39 posts)→

Original post →

More from AGI Musings

AGI Musings channel →