The Ahmed body (Ahmed et al., 1984, SAE Paper 840300) is certainly one of the most frequently computed benchmarks for studying car aerodynamics. Actually, the car aerodynamics is here reduced on the flow separation at the rear slant and the subsequent wake region. Two most interesting Ahmed body configurations are those characterized by the slant angles 25 and 35 degrees. The main difference between these two cases is that the flow detaches and reattaches at the slanted surface for the lower slant angle value, while for the 35-slant-angle the flow detaches and forms a fairly large single recirculation region at the entire slanted surface, turning into a wake region behind the body. Whereas the underlying mean flow structure of the 35-slant-angle case can be correctly returned by applying turbulence models in the (steady or conventional unsteady) RANS (Reynolds-Averaged Navier Stokes) framework, the RANS computations of the 25-slant-angle configuration result in a completely wrong flow topology, resembling actually that pertinent to the 35-slant-angle case (see e.g., Guilmineau, JWE-and-IA 2008). The reason for such an incorrect outcome lies primarily in an inadequately low turbulence activity within the separated shear layer, implying a low intensity momentum exchange between the main stream and separated region. This is the expected consequence of the physical rationale of the RANS method, which is, as such, not capable of capturing the large-scale eddy structures with a broader spectrum and the bulk flow unsteadiness – as encountered in the flows involving separation. Correct capturing of the spectral dynamics, typical for the separated shear layer region, is possible only by applying an LES (Large-Eddy Simulation) or relevant model schemes (see e.g., Krajnovic and Davidson, ASME JFE, 2005). However, it implies an unsteady, time-accurate flow simulation in conjunction with an appropriately fine grid resolution, which is not easily affordable in the case of practically-oriented computational activities. In the present paper however, a turbulence model extension is presented (Jakirlic and Maduta, AIAA J. 54(5), 2016) which makes it possible to obtain correct results of the 25-slant-angle case even when applying a steady RANS method.