Computational Fluid Dynamics (CFD) analysis of complex external flows is an important component of road vehicle design and development. The Reynolds Averaged Navier Stokes (RANS) turbulence modeling is widely used in the automotive industry for simulating these complex flows, because of its short turnover time and cost-effectiveness. Existing literature shows that the popular two-equation turbulence models like, the SST k − ω turbulence model, fail to produce good flow predictions when applied to a simplified car geometry, such as the Ahmed body, which exhibits flow features similar to real automotive flows. The present study improves the prediction capabilities of the SST k−ω turbulence model by tuning a few selected turbulence model closure coefficients for the 25° rear slant angle (φ = 25°) Ahmed body. This involves studying individual effects of these closure coefficients and then formulating a combination of these parameters that yields the best correlation with the experimental data. The best combination thus attained is then applied to scale resolved simulations such as the Improved Delayed Detached Eddy simulations (IDDES) for improved flow predictions. The tuning of closure coefficients was further applied to Ahmed body geometry with different slant angles (φ = 20° and φ = 35°). The present study reveals that a combination of the closure coefficients can be obtained that can lead to very well correlated force and flow predictions for each of the Ahmed body configurations. However, this tuned combination is not universal, and each slant angle requires a different combination of model closure coefficients. This, in turn, questions the validity of equilibrium assumptions used in postulating the turbulence model transport equations for flows with massive separation and reattachment.