Benchmarking numerical models is essential for validating their accuracy and ensuring consistency across simulation platforms. This study presents a comparative benchmark analysis of two widely used Large Eddy Simulation (LES) codes, AMR-WIND and NREL SOWFA-6, focusing on wind turbine rotor performance, wake dynamics, and atmospheric boundary layer (ABL) representation. The evaluation includes an actuator line model (ALM)-based uniform inflow wind turbine simulation and ABL precursors under neutral and unstable conditions. The uniform inflow wake analysis examined differences in wind turbine induction and wake development between the two codes. Additionally, neutral and unstable atmospheric boundary layer precursors were generated for an offshore environment and compared. Results indicate a difference in wake breakdown location between the codes (one contributing factor was the difference in numerical schemes used for the advection terms.) The number of actuator points required for smooth velocity distribution across the rotor was higher for SOWFA-6 than AMR-WIND. In ABL precursors, time-averaged flow fields showed strong agreement, though minor discrepancies in turbulence were observed, particularly in unstable conditions, affecting coherence analysis. The energy distribution across wavenumbers showed a good match between the codes, with slight discrepancies observed in the large and small wavenumber regions. The cutoff wavenumber was found to be similar for both codes. Lateral and vertical coherence at small and large separations were in close agreement for the neutral ABL. However, in the unstable ABL, notable differences in coherence were observed between the codes for separations greater than 40 m.