LeMat-GenBench: A Unified Evaluation Framework for Crystal Generative Models

TL;DR

LeMat-GenBench provides a standardized evaluation framework and benchmark suite for crystalline material generative models, facilitating fair comparison and guiding future development in materials discovery.

Contribution

It introduces a unified benchmark with evaluation metrics, an open-source suite, and a public leaderboard for crystalline generative models, addressing the lack of standardization.

Findings

Stability correlates with reduced novelty and diversity.

No model outperforms others across all metrics.

Benchmarking reveals trade-offs among model properties.

Abstract

Generative machine learning (ML) models hold great promise for accelerating materials discovery through the inverse design of inorganic crystals, enabling an unprecedented exploration of chemical space. Yet, the lack of standardized evaluation frameworks makes it challenging to evaluate, compare, and further develop these ML models meaningfully. In this work, we introduce LeMat-GenBench, a unified benchmark for generative models of crystalline materials, supported by a set of evaluation metrics designed to better inform model development and downstream applications. We release both an open-source evaluation suite and a public leaderboard on Hugging Face, and benchmark 12 recent generative models. Results reveal that an increase in stability leads to a decrease in novelty and diversity on average, with no model excelling across all dimensions. Altogether, LeMat-GenBench establishes a reproducible and extensible foundation for fair model comparison and aims to guide the development of more reliable, discovery-oriented generative models for crystalline materials.