In this study, we assess uncertainties of recent digital elevation models (DEMs) in North America and review their coverage, resolution, format, and data used to create them. Uncertainties are assessed in terms of their internal error and external accuracy. The internal error of the DEMs is usually expressed as nominal accuracy by the model providers and the external accuracy is estimated by comparing the models with independent observations. We use the benchmarks with GPS/GNSS observations in Canada, the USA, and Mexico as control points to estimate the external accuracy. Orthometric heights at the control points are computed using the geoid heights of the (latest) gravimetric geoid model over North America and geodetic heights by GPS/GNSS observations. Then the orthometric heights for the finest resolution of each DEM are extracted at the control points for computing the external accuracy. The DEMs of interest in this study include the Shuttle Radar Topography Model V4.0 (SRTM), Altimeter Corrected Elevation (ACE), Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), Multi-Error-Removed Improved-Terrain (MERIT), and TerraSAR-X add-on for Digital Elevation Measurements (TanDEM-x).