Atomically precise metal nanoclustersMetal nanoclusters have recently risen in popularity within the field of nanomaterialsNanomaterials due to having unique molecule-like properties. At such a small scale, important chemical properties, such as the binding energies and the HOMO–LUMO gapHOMO-LUMO gap of such species are heavily dependent on their geometry and the interaction between the atoms. Quantum calculation methods, namely density functional theoryDensity Functional Theory (DFT), have been used to study the electronic structure of these nanoclusters. However, nanoclusters occupy a vast chemical space due to their diverse size, composition and geometry, which quantum calculation cannot explore in a time-efficient manner. InsteadMachine learning (ML), machine learningMachine learning (ML)-based methods, namely graph neural networksGraph neural network have recently proven to achieve near-quantum calculation levels of accuracy while delivering high throughput, making them a more suitable candidate for future nanocluster design and screening tasks. Herein, we investigate the performance of conventional chemical featurization methods along with graph neural networksGraph neural network on a dataset of metal nanoclusters.