Previously, the Penn State/National Center for Atmospheric Research (PSU/NCAR) Fifth Generation Mesoscale Model (MM5) has been successfully adapted for the polar regions. The optimized model, known as Polar MM5, has been used for many applications in the Arctic and Antarctic regions. MM5, however, is reaching the end of its prime use years. Its replacement model, the Weather Research and Forecasting model (WRF) has been developed as a collaboration by several working groups and agencies and is now operational. The new model's physics is also being adapted for the unique environments of the Arctic and Antarctic. The polar-optimized WRF will be applied to several tasks including the Arctic System Reanalysis, a planned high resolution reanalysis including atmospheric, oceanic, and and sea ice components. The new model will also serve as the base model for future Antarctic Mesoscale Prediction System (AMPS) operational forecasts. To evaluated the performance of WRF prior to the polar optimization, version 1.3 of WRF has been run for July and December 2002 on an Arctic grid with 60 km resolution including several Alaskan and Canadian river basins in the domain. It has also been run for the same months with 40 km resolution on a North Atlantic grid centered over Greenland. The simulations show that the synoptic-scale weather patterns over the Arctic and the synoptic-scale weather variations over Greenland are reasonably captured by the early version of WRF. There are significant biases, however, in the boundary layer fields during December 2002 for Greenland. In general, the boundary layer temperature is too warm for December and the surface wind speed is underestimated. Improvements are clearly needed to the model physics. The origins of these biases are being diagnosed.