The first isotope-enabled general circulation model (GCM) simulations of the Pliocene are used to discuss the interpretation of 18 O measurements for a warm climate. The model suggests that spatial patterns of Pliocene ocean surface 18 O ( 18 O sw ) were similar to those of the preindustrial period; however, Arctic and coastal regions were relatively depleted, while South Atlantic and Mediterranean regions were relatively enriched. Modeled 18 O sw anomalies are closely related to modeled salinity anomalies, which supports using 18 O sw as a paleosalinity proxy. Modeled Pliocene precipitation 18 O ( 18 O p ) was enriched relative to the preindustrial values (but with depletion of < 2‰ over some tropical regions). While usually modest (<4‰), the enrichment can reach 25‰ over ice sheet regions. In the tropics 18 O p anomalies are related to precipitation amount anomalies, although there is usually a spatial offset between the two. This offset suggests that the location of precipitation change is more uncertain than the amplitude when interpreting 18 O p . At high latitudes 18 O p anomalies relate to temperature anomalies; however, the relationship is neither linear nor spatially coincident: a large 18 O p signal does not always translate to a large temperature signal. These results suggest that isotope modeling can lead to enhanced synergy between climate models and climate proxy data. The model can relate proxy data to climate in a physically based way even when the relationship is complex and nonlocal. The 18 O-climate relationships, identified here from a GCM, could not be determined from transfer functions or simple models.