Periodic Labs' Neon, trained on 1,300 H200s, beats GPT-6 Astra on its science benchmark

_AndrewZhao · x · 2026-09-16

Periodic Labs (co-founded by ex-OpenAI's Liam Fedus) has built high-throughput materials labs in Menlo Park targeting superconductors, magnets and semiconductor materials, closing the loop between experiments and models: labs generate fresh data, models learn from it and then pick the next experiments. Experiments now run continuously, up from early manual runs. Using only 1,300 H200 GPUs plus months of proprietary lab data, they mid-trained and RL'd an open-source model into 'Neon,' which surpasses GPT-6 Astra on their analysis benchmark.

Related event: Periodic Labs Trains Materials Model Beating GPT-6 with Just 1,300 H200s(31 posts)→

Original post →

More from Infra

Infra channel →