A high-throughput materials science lab and model-loop system set up in Menlo Park produces fresh experimental data, which were used to further train an open-source model via mid-training and reinforcement learning; using 1,300 H200 GPUs, the model named Neon exceeded GPT-6 Astra on an analysis benchmark. The project focuses on hard problems in superconductors, magnets and semiconductor materials, and speeds up experiment–model interaction.
AI-generated text
Neon: Menlo Park high-throughput materials science lab and open model to surpass GPT-6 Astra
A high-throughput materials science lab and model-loop system set up in Menlo Park produces fresh experimental data, which were used to further train an open-source model via mid-training and reinforcement learning; using 1,300 H200 GPUs, the model named Neon exceeded GPT-6 Astra on an analysis benchmark.
