The previous interglacial (Marine isotope stage 5e, peaking at 125 kyr BP) is usually assumed to be a good analogue for our present-day climate. Moreover, as our present interglacial, the Holocene, is already almost 12 kyr long, paleoclimatologists were inclined to predict a quite close entrance into the next ice age. Simulations using the 2.5-D climate model of Louvain-la- Neuve show however that our interglacial will most probably last much longer than any previous one, even without human intervention (Berger and Loutre, 2002). It was suggested that this is related to the shape of the Earth’s orbit around the Sun which will be almost circular over the next tens of thousands of years. A circular orbit means indeed that the latitudinal and seasonal distributions of the energy received from the Sun will not vary anymore, a situation which will give more weight to the other forcings, in particular to the forcing by the greenhouse gases. As the eccentricity variation is primarily related to the 400-kyr cycle, the best and closer analogue for such a forcing is definitively Marine isotopic Stage 11 (MIS-11) some 400 kyr ago and not MIS-5e. Simulations of this MIS-11 interglacial under greenhouse gas and astronomical forcings led indeed also to an anomalous length, a prediction which was later confirmed by the EPICA ice-core record. Such a relationship between CO2 and climate is at the basis of the claim by Ruddiman (2003) that the impact of human activities on climate might have already started 10,000 years ago, preventing our climate to have already entered into glaciation. Sensitivity experiments with the LLN model tend to rule out Ruddiman’s hypothesis because glacial inception in this model would require CO2 concentration below 240 ppmv during the Holocene. However, honest estimates of model uncertainties incline us to leave the question open (Crucifix and Berger, 2006). Our results show not only that the Holocene might last particularly long, but also that the sensitivity of our climate system to the greenhouse gas forcing might be exacerbated. It is therefore not surprising that, in our model at least, there is a threshold in the greenhouse gas concentration of about 700 ppmv beyond which the Greenland ice sheet melts in about 5000 years and does not recover before a few tens of thousands of years. All these results confirm that the CO2 concentration in the atmosphere plays an important role in shaping the long-term climatic variations and that a detailed reconstruction of the interglacials from land, deep-sea and ice records is urgently needed.