Periodic Labs midtrains and RLs a trillion-param LLM to analyze superconductor lab data, beating Astra and Fable

vwxyzjn · x · 2026-09-16

Periodic Labs has mid-trained and RL'd a trillion-parameter LLM to analyze experimental data from its superconductor lab. The team claims it outperforms Astra and Fable on this task and that their scientists love using it. The company published its first research report alongside the announcement, showcasing its experiment-to-model data loop.

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

Original post →

More from Research

Research channel →