# UniFFBench: Evaluating Universal Machine Learning Force Fields Against Experimental Measurements

arXiv: https://arxiv.org/abs/2508.05762 · HTML: https://mlsanity.com/paper/2508.05762/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Builds on 14 of its 47 references that are in the map.
- First posted on arXiv 7 Aug 2025.

## Builds on (most cited first, 14 of 14)

- [Learning Local Equivariant Representations for Large-Scale Atomistic Dynamics](https://mlsanity.com/paper/2204.05249/) (2022)
- [The Open Catalyst 2020 (OC20) Dataset and Community Challenges](https://mlsanity.com/paper/2010.09990/) (2020)
- [UMA: A Family of Universal Models for Atoms](https://mlsanity.com/paper/2506.23971/) (2025)
- [The Open Catalyst 2022 (OC22) Dataset and Challenges for Oxide Electrocatalysts](https://mlsanity.com/paper/2206.08917/) (2022)
- [Forces are not Enough: Benchmark and Critical Evaluation for Machine Learning Force Fields with Molecular Simulations](https://mlsanity.com/paper/2210.07237/) (2022)
- [Learning Smooth and Expressive Interatomic Potentials for Physical Property Prediction](https://mlsanity.com/paper/2502.12147/) (2025)
- [A Hitchhiker's Guide to Geometric GNNs for 3D Atomic Systems](https://mlsanity.com/paper/2312.07511/) (2023)
- [Orb: A Fast, Scalable Neural Network Potential](https://mlsanity.com/paper/2410.22570/) (2024)
- [Probing the limitations of multimodal language models for chemistry and materials research](https://mlsanity.com/paper/2411.16955/) (2024)
- [Learning neural network potentials from experimental data via Differentiable Trajectory Reweighting](https://mlsanity.com/paper/2106.01138/) (2021)
- [EGraFFBench: Evaluation of Equivariant Graph Neural Network Force Fields for Atomistic Simulations](https://mlsanity.com/paper/2310.02428/) (2023)
- [The Open MatSci ML Toolkit: A Flexible Framework for Machine Learning in Materials Science](https://mlsanity.com/paper/2210.17484/) (2022)
- [MatSciML: A Broad, Multi-Task Benchmark for Solid-State Materials Modeling](https://mlsanity.com/paper/2309.05934/) (2023)
- [Energy & Force Regression on DFT Trajectories is Not Enough for Universal Machine Learning Interatomic Potentials](https://mlsanity.com/paper/2502.03660/) (2025)

## Built on by (most cited first, 6 of 6)

- [Elign: Equivariant Diffusion Model Alignment from Foundational Machine Learning Force Fields](https://mlsanity.com/paper/2601.21985/) (2026)
- [Generalization of Long-Range Machine Learning Potentials in Complex Chemical Spaces](https://mlsanity.com/paper/2512.10989/) (2025)
- [Refining Machine Learning Potentials through Thermodynamic Theory of Phase Transitions](https://mlsanity.com/paper/2512.03974/) (2025)
- [GFFMERGE: Efficient Merging of Graph Neural Force Fields and Beyond](https://mlsanity.com/paper/2606.03232/) (2026)
- [Empowering Polymeric Materials Discovery by Artificial Intelligence](https://mlsanity.com/paper/2606.20753/) (2026)
- [On the Theoretical Limits of Learning Unconstrained Interatomic Potentials](https://mlsanity.com/paper/2610.15078/) (2026)

## Abstract (arXiv)

Universal machine learning force fields (UMLFFs) promise to revolutionize materials science by enabling rapid atomistic simulations across the periodic table. However, their evaluation has been limited to computational benchmarks that may not reflect real-world performance. We introduce UniFFBench, a comprehensive evaluation framework featuring the MinX dataset -- a diverse collection of 1,500+ mineral systems spanning 85 elements, extreme thermodynamic conditions (0--5000 K, 0--1000 GPa), and structural complexity, including partial occupancy and disorder. This diversity, combined with experimental reference values for validation, enables assessment of UMLFF generalization across chemical space and conditions substantially beyond typical training scenarios. Our systematic evaluation of six state-of-the-art UMLFFs reveals a substantial ``reality gap'': models achieving impressive performance on computational benchmarks often fail when confronted with experimental complexity. Even the best-performing models exhibit higher density prediction error than the threshold required for practical applications. We observe disconnects between simulation stability and mechanical property accuracy, with prediction errors correlating with training data representation rather than the modeling method.