Two versions of a coupled atmosphere-ocean general circulation model (GCM) with different climate sensitivities are tested on global cooling following the Pinatubo volcanic eruption to investigate the validity of high climate sensitivities. The higher-sensitivity version, with climate sensitivity of 6.3 K for doubled CO2 forcing, overestimates cooling due to the volcanic eruption, whereas the lower-sensitivity version (4.0 K) produces results consistent with observations. A simple scheme for climate feedback analysis is devised and it is found that the difference between the two versions is attributed to cloud-albedo feedback. This validation method is expected to provide additional constraints on climate sensitivity and possibly lead to reduced uncertainties in climate prediction.