A foundation model for atomistic materials chemistry

I Batatia and P Benner and Y Chiang and AM Elena and DP Kovács and J Riebesell and XR Advincula and M Asta and M Avaylon and WJ Baldwin and F Berger and N Bernstein and A Bhowmik and F Bigi and SM Blau and V Carare and M Ceriotti and S Chong and JP Darby and S De and F Della Pia and VL Deringer and R Elijosius and Z El-Machachi and E Fako and F Falcioni and AC Ferrari and JLA Gardner and MJ Gawkowski and A Genreith-Schriever and J George and REA Goodall and J Grandel and CP Grey and P Grigorev and S Han and W Handley and HH Heenen and K Hermansson and CH Ho and S Hofmann and C Holm and J Jaafar and KS Jakob and H Jung and V Kapil and AD Kaplan and N Karimitari and JR Kermode and P Kourtis and N Kroupa and J Kullgren and MC Kuner and D Kuryla and G Liepuoniute and C Lin and JT Margraf and IB Magdau and A Michaelides and JH Moore and AA Naik and SP Niblett and SW Norwood and N O'Neill and C Ortner and KA Persson and K Reuter and AS Rosen and LAM Rosset and LL Schaaf and C Schran and BX Shi and E Sivonxay and TK Stenczel and C Sutton and V Svahn and TD Swinburne and J Tilly and C van der Oord and S Vargas and E Varga-Umbrich and T Vegge and M Vondrák and YS Wang and WC Witt and T Wolf and F Zills and G Csányi, JOURNAL OF CHEMICAL PHYSICS, 163, 184110 (2025).

DOI: 10.1063/5.0297006

Atomistic simulations of matter, especially those that leverage first- principles (ab initio) electronic structure theory, provide a microscopic view of the world, underpinning much of our understanding of chemistry and materials science. Over the last decade or so, machine- learned force fields have transformed atomistic modeling by enabling simulations of ab initio quality over unprecedented time and length scales. However, early machine-learning (ML) force fields have largely been limited by (i) the substantial computational and human effort required to develop and validate potentials for each particular system of interest and (ii) a general lack of transferability from one chemical system to the next. Here, we show that it is possible to create a general-purpose atomistic ML model, trained on a public dataset of moderate size, that is capable of running stable molecular dynamics for a wide range of molecules and materials. We demonstrate the power of the MACE-MP-0 model-and its qualitative and at times quantitative accuracy- on a diverse set of problems in the physical sciences, including properties of solids, liquids, gases, chemical reactions, interfaces, and even the dynamics of a small protein. The model can be applied out of the box as a starting or "foundation" model for any atomistic system of interest and, when desired, can be fine-tuned on just a handful of application-specific data points to reach ab initio accuracy. Establishing that a stable force-field model can cover almost all materials changes atomistic modeling in a fundamental way: experienced users obtain reliable results much faster, and beginners face a lower barrier to entry. Foundation models thus represent a step toward democratizing the revolution in atomic-scale modeling that has been brought about by ML force fields.

Return to Publications page