Foundation models for materials modeling are advancing quickly, but their training remains expensive, often placing state-of-the-art methods out of reach for many research groups. We introduce Nequix, a compact E(3)-equivariant potential that pairs a simplified NequIP design with modern training practices, including equivariant root-mean-square layer normalization and the Muon optimizer, to retain accuracy while substantially reducing compute requirements. Nequix has 700K parameters and was trained in 100 A100 GPU-hours. On the Matbench-Discovery and MDR Phonon benchmarks, Nequix ranks third overall while requiring a 20 times lower training cost than most other methods, and it delivers two orders of magnitude faster inference speed than the current top-ranked model. We release model weights and fully reproducible codebase at https://github.com/atomicarchitects/nequix.
Cite
@article{arxiv.2508.16067,
title = {Training a Foundation Model for Materials on a Budget},
author = {Teddy Koker and Mit Kotak and Tess Smidt},
journal= {arXiv preprint arXiv:2508.16067},
year = {2025}
}
Comments
To appear at the NeurIPS 2025 Workshop on AI for Accelerated Materials Discovery