NEP89: universal neuroevolution potential for inorganic and organic materials across 89 elements.

NEP89 delivers high-accuracy, computationally efficient atomistic simulations across 89 elements, enabling large-scale studies previously limited by resource demands.

• Why it matters: Current machine-learned potentials are often material-specific or computationally intensive, restricting their widespread application in diverse atomistic modeling scenarios.
• What they did: The authors developed NEP89, a foundation model based on neuroevolution potential architecture, trained on a curated dataset covering inorganic and organic materials, achieving broad applicability.
• The result: NEP89 is three to four orders of magnitude more efficient than comparable models, facilitating large-scale simulations and rapid fine-tuning for various material properties and complex systems.