ReMODEL: A Large-Scale, Multi-Fidelity Experimental Dataset for AI-Driven Metal-Organic Framework Discovery

We propose ReMODEL, a large-scale, open dataset designed to accelerate the use of AI in materials discovery. Much like how the Protein Data Bank enabled breakthroughs such as AlphaFold in biology, ReMODEL aims to do the same for the discovery of crystalline materials known as metal–organic frameworks (MOFs). By systematically collecting and sharing thousands of experimental outcomes—including both successes and failures—ReMODEL will provide a rich foundation for training AI models to predict, design, and understand new materials. The project will support advances in clean energy, climate solutions, and scientific understanding, while promoting open science and collaboration across disciplines.