The Last Interglacial (LIG, 130–115 kiloyear (kyr) before present (BP)) and the Holocene (10–0 kyr BP) provide a good test bed for climate models that are used for future climate projections, as the climatic forcings during these interglacial periods are well constrained. The LIG is characterized by a strong seasonal insolation forcing as compared to the present, that is driven by variations in the Earth's astronomical configuration. During the LIG, the northern high latitudes experienced higher temperatures than those of the late Holocene, as indicated by proxy records and modelling studies, and the Greenland Ice Sheet (GIS) was notably reduced. However, the impact of a reduced GIS on the global climate has not yet been well constrained. In this study, the contribution of the GIS to LIG warmth is quantified by performing various sensitivity studies, employing the Community Earth System Models (COSMOS). The focus is set on height and extent of the GIS. In order to assess the effects of insolation changes over time on temperature evolution and seasonality, and for a comparison of LIG climate with the current interglacial, transient simulations, covering the whole LIG and Holocene, are performed. The resulting surface temperature fields are analysed, and the contribution of different forcings to LIG warmth is separated. It is found that strong Northern Hemisphere warming is mainly caused by increased summer insolation. Reducing height and extent of the GIS leads to an additional warming of several degrees Celsius in the northern and southern high latitudes during local winter. In order to evaluate the performance of the COSMOS LIG simulations, the simulated surface temperature anomalies are compared to marine and terrestrial proxy-based LIG temperature anomalies. It is found that model results are in good agreement with proxy records with respect to the spatial pattern of the temperature change, but they underestimate the reconstructed temperatures. This mismatch between model and data is reduced by taking into account potential seasonal biases of the proxy records and by considering uncertainties in the dating of the proxy records for the LIG thermal maximum. Respective seasonal biases and uncertainty in the dating are estimated from the transient simulations performed with COSMOS. However, the COSMOS LIG simulations are not able to reproduce the full magnitude of temperature changes indicated by the proxies. A similar underestimation of proxybased temperatures by the COSMOS is found also for the Holocene period. In order to test whether the model-data mismatch is model-dependent, time slice simulations from the Paleoclimate Modelling Intercomparison Project (PMIP), representing LIG and Holocene temperatures, are as well compared to the proxy reconstructions. This comparison indicates an underestimation of the proxy reconstructions by all considered models. These results indicate that simulated and reconstructed temperature changes are, to a large degree, only qualitatively comparable, suggesting a potential misinterpretation of the proxy records, and/or deficits of the models, such as model sensitivity to orbital forcing.