Measurements of turbulent fluxes, clouds, radiation, and profiles of mean meteorological parameters, obtained over an ice floe in the central Arctic Ocean during the Arctic Ocean Experiment 2001, are used to evaluate the performance of U.K. Met Office Unified Model (MetUM) and Coupled Ocean/Atmosphere Mesoscale Prediction System (COAMPS) in the lower atmosphere during late summer. Both the latest version of the MetUM and the version operational in 2001 are used in the comparison to gain an insight as to whether updates to the model have improved its performance over the Arctic region. As with previous model evaluations over the Arctic, the pressure, humidity, and wind fields are satisfactorily represented in all three models. The older version of the MetUM underpredicts the occurrence of low-level Arctic clouds, and the liquid and ice cloud water partitioning is inaccurate compared to observations made during SHEBA. In the newer version, simulated ice and liquid water paths are improved, but the occurrence of low-level clouds are overpredicted. Both versions overestimate the amount of radiative heat absorbed at the surface, leading to a significant feedback of errors involving the surface albedo, which causes a large positive bias the surface temperature. Cloud forcing in COAMPS produces similar biases in the downwelling shortwave and longwave radiation fluxes to those produced by UM(G25). The surface albedo parameterization is, however, more realistic, and thus, the total heat flux and surface temperature are more accurate for the majority of the observation period.