We analyze data from seven continuous Global Positioning System ( GPS) receivers and one tide gauge, all located along the edge of the Greenland ice sheet, to determine vertical uplift rates. We compare our results with predictions based on the ICE-5G deglaciation model of Peltier ( 2004). Results from the GPS receiver at Kellyville (-1.2 +/- 1.1 mm/a) and from the tide gauge at Nuuk (-2.2 +/- 1.3 mm/ a), indicate that ICE-5G overestimates the subsidence rates at those locations by 2.1 and 1.1 mm/ a, respectively. Kellyville and Nuuk are located along the southwestern margin of the Greenland ice sheet, and the observed negative uplift rates are consistent with independent evidence that the ice margin along the southwestern edge readvanced during the last similar to 8 ka to its current position. The ICE-5G glaciation-deglaciation history includes a readvance between the latitudes of 62 degrees N and 72 degrees N. The GPS measurements suggest the ICE-5G readvance may be too large or mistimed. Our GPS results at Qaqortoq, located at the southern tip of Greenland, suggest a secular subsidence rate of -0.3 +/- 1.1 mm/ a, while ICE-5G predicts an uplift rate of 1.0 mm/ a. ICE-5G assumes no ice sheet readvance in south Greenland, including no readvance of the Qassimiut lobe. The difference of 1.3 +/- 1.1 mm/ a can tentatively be explained as due to a similar to 33 km readvance of the Qassimiut lobe during the last similar to 3 ka. For the other GPS sites, the observed/ predicted uplift rates are 3.6 +/- 1.1/-0.1 mm/ a at Thule, 0.0 +/- 1.1/2.0mm/a at Scoresbysund, and -0.4 +/- 1.1/-1.7mm/a at Kulusuk. For Thule, Kulusuk, and Scoresbysund the differences between the observed and predicted rates are on the order of 1.3 - 3.7 mm/ a, though with opposite signs, and indicate that ICE-5G does not exactly reproduce the correct rebound signal at those locations.