We report results of a three-year intercomparison experiment between the WHOI Radiocarbon Laboratory (now at University of California, Irvine) and the NSF-University of Arizona AMS Laboratory. The purpose of this study was to compare high-precision measurements of samples obtained routinely using gas proportional counting techniques with results obtained using AMS techniques. Three sets of annually banded, modern coral samples were used for the intercomparison. Each sample was acidified to CO2 at WHOI and split into two fractions. The larger fraction was converted to acetylene gas and counted at WHOI in quartz, gas-proportional beta counters. The smaller fractions were converted to graphite and analyzed for 14C using AMS techniques at the University of Arizona. Results of the three sample sets are presented. Except for a single outlier, the data from the two laboratories are in good agreement. Of the 13 samples in the third set of the intercomparison, for which a new high-intensity ion source was in operation at the Arizona AMS laboratory, agreement of results is excellent. This finding indicates that measurements made with precisions of < 3% are reproducible at both laboratories.