Climate is often depicted as globally warm during the middle part of the Cretaceous. However, major tectonic events, such as the break-up of Gondwana, or sea level changes inducing the flooding of vast continental areas, occurred from 120 to 95 Ma resulting in drastic climatic re-organisations. In order to investigate the impact of plate tectonics and sea level change on climate, we performed climatic simulations during the Cenomanian (95 Ma) and Aptian (120 Ma) using an atmospheric general circulation model. Problems with models lie in our knowledge of a few key factors. For examples, the middle Cretaceous sea surface temperature gradient from equator to pole remains an open question and it is not clear yet how flat it was. Two sensitivity experiments for each period prescribing a flatter SST gradient based on oxygen isotopic data have been done in order to test its climatic consequences. We discuss separately the effect of palaeogeography and meridional SST gradient impacts. In the northern hemisphere, the flooding of large continental areas due to sea level rise leads to a more equable climate in the mid-latitudes. However the impact of epicontinental seas varies seasonally from site to site due to local feedbacks. In the southern hemisphere, the breakup of Gondwana deeply affects the climate of this region, moderating the seasonal thermal variations. In addition, we test the impact of a flatter SST gradient on climate and we show that simulations with flat SST gradients compare well with field data. We analyse the remaining model/data discrepancies in terms of missing feedbacks (CO2, vegetation) but also of new measurements that enable to infer a better meridional variation of sea surface temperature.