Researchers have developed the Anchored-Branched Steady-state Wind Flow Transformer (AB-SWIFT), a metamodel designed to simulate 3D atmospheric flow in urban environments, addressing the need for efficient air flow modeling at a local scale1. This innovation aims to bypass the costly Computational Fluid Dynamics (CFD) computations required for such modeling. Deep learning surrogate models have shown promise as alternatives, but they struggle to adapt to the high variability of urban geometries. The AB-SWIFT model is intended to improve upon these limitations, providing a more effective solution for applications like pollutant dispersion modeling and wind farm modeling. By leveraging this metamodel, practitioners can better understand and predict air flow patterns in complex urban environments. This matters to urban planners and environmental modelers because accurate air flow modeling is crucial for mitigating the effects of pollution and optimizing wind energy production, and AB-SWIFT has the potential to significantly enhance their capabilities.