Urban design requires accurate prediction of wind, gusts, and contaminant dispersion to create sustainable cities. While computational fluid dynamics (CFD) models like large-eddy simulation (LES) provide accurate predictions, they are computationally expensive. Steady Reynolds-averaged Navier-Stokes (SRANS) is fast but inaccurate due to inherent limitations. This study proposes a two-stage CFD-graph neural network (GNN) framework for efficient steady-state prediction. A CFD solver provides the initial state employing SRANS with the k-ε model, and then a GNN reconstructs the CFD outputs towards a high-fidelity target in a single inference step, bypassing SRANS flaws. CFD-GNN employing a coarse mesh achieved considerable agreement with the fine LES result and was thousands of times faster. On the same coarse mesh, CFD-GNN was seven times faster than a tightly converged SRANS and 280 times faster than an under-resolved LES.