We present a coupling approach to and the first results of the GRISLI ice-sheet model within the iLOVECLIM-coupled climate model. The climate component is a relatively low-resolution earth system model of intermediate complexity, well suited for long-term integrations and thus for coupled climate-cryosphere studies. We describe the coupling procedure with emphasis on the downscaling scheme and the methods to compute the snow fraction from total precipitation fields. We then present results for the Greenland ice sheet under pre-industrial climate conditions at the end of a 14 000 yr long integration. The simulated ice sheet presents too large a thickness in its central part owing to the overestimation of precipitation in the atmospheric component. We find that including downscaling procedures for temperature improves the temperature distributions over Greenland for both the summer and annual means. We also find an ice-sheet areal extent that is overestimated with respect to the observed Greenland ice sheet. © Author(s) 2014.