Detached-Eddy Simulation (DES) is applied to prediction of the super-critical flow around a circular cylinder. One of the primary aims is to assess a new DES version developed by Spalart et al.1 against results obtained using the baseline method. In the new version of the technique, known as Delayed Detached Eddy Simulation (DDES), the turbulent length scale is determined using information from the eddy viscosity Held, in addition to the wall distance and grid spacing. Computations are performed at Reynolds numbers, Re, based on the freestream velocity and cylinder diameter of 1.4 × 105 and 8 × 106, with the lower Re predictions assessed against previous simulations and the higher Re assessed against experimental measurements. Flow visualizations show that there is comparable eddy content resolved using the baseline and new DES versions. At Re = 1.4 × 105, predictions of the drag coefficient, separation angle, and pressure distribution are in good agreement with the fully turbulent solutions of Travin et al.2 and Hansen and Forsythe.3 Predictions at Re = 8 × 106 are obtained using three grids with the coarsest mesh having 1.47 × 106 cells and the finest grid having 9.83 × 106 cells. The force histories and averaged force coefficients obtained using both models are in good agreement. Predictions of the pressure coefficient using the baseline and new DES versions are in the range of experimental measurements.