Description: Here we present the sensitivity experiments are conducted with the latest GRISLI version 2.0 (Quiquet et al. 2018), to model changes of the Antarctic ice sheet between 100 ka and 400 ka. GRISLI 2.0 is is a large-scale three-dimensional thermomechanical ice sheet model and the newest version of the Grenoble Ice Sheet and Land Ice model (Ritz et al. 2001). The model combines an inland ice model with an ice shelf model, extended to the case of ice streams considered as dragging ice shelves. The latest release includes a better representation of grounding line migration and a sub-glacial hydrology model. The model uses finite differences on a Cartesian grid at 5 to 40 km resolution depending on the application. Here we use a 40 km grid. The spatial coverage of the model experiments is latitude: -90/-54.6602 ; longitude: -180/180. For the selection of the ice sheet model parameters, we used the calibration performed in Quiquet et al. (2018) using the ensemble member labeled AN40T213 in Quiquet et al. (2018). This ensemble member uses the formulation of Tsai et al. (2015) for the imposed flux at the grounding line. The model setup used to perform the transient paleo ice sheet simulations (boundary conditions and forcings) is identical to that used by Quiquet et al. (2018). The model is forced by near-surface air temperatures over Antarctica deduced from the EDC δD record (Jouzel et al. 2007). The GRISLI simulations are performed prescribing three different oceanic forcing indexes: (i) derived from the North Atlantic ODP 980 benthic temperature record (Quiquet et al., 2018; Waelbroeck et al., 2002); (ii) derived from the stacked deep-sea benthic oxygen isotope record (LR04)(Golledge et al., 2014; Lisiecki & Raymo, 2005); and (iii) derived from the EDC δD record (Blasco et al., 2019; Golledge et al., 2014; Jouzel et al., 2007). The outputs of the model consist in the Antarctic ice sheet surface elevation changes, bedrock elevation, and ice thickness variations over the past 400 ka with a 1ka temporal resolution. In addition, we calculated the evolution of the ice volume contained in the Wilkes Basin during the past 400 ka. We performed 27 sensitivity tests in which we changed the initial state of the Antarctic ice sheet (i.e. glacial, interglacial or deglaciated state) at 400 ka. In addition, given that the Southern Ocean forcing represents a large source of uncertainty, for all the simulations at the glacial, interglacial and deglaciated Antarctic initial state, we perform three sensitivity simulations changing the oceanic forcing: the original oceanic and the original forcing increased by 5% and by 10%.