Climate sensitivity and feedbacks are basic and important metrics to a climate system. They determine how large surface air temperature will increase under CO2 forcing ultimately, which is essential for carbon reduction policies to achieve a specific warming target. In this study, these metrics are analyzed in a climate system model newly developed by the Chinese Academy of Meteorological Sciences (CAMS-CSM) and compared with multi-model results from the Coupled Model Comparison Project phase 5 (CMIP5). Based on two idealized CO2 forcing scenarios, i.e., abruptly quadrupled CO2 and CO2 increasing 1% per year, the equilibrium climate sensitivity (ECS) and transient climate response (TCR) in CAMS-CSM are estimated to be about 2.27 and 1.88 K, respectively. The ECS is near the lower bound of CMIP5 models whereas the TCR is closer to the multi-model ensemble mean (MME) of CMIP5 due to compensation of a relatively low ocean heat uptake (OHU) efficiency. The low ECS is caused by an unusually negative climate feedback in CAMS-CSM, which is attributed to cloud shortwave feedback (λSWCL) over the tropical Indo-Pacific Ocean. The CMIP5 ensemble shows that more negative λSWCL is related to larger increase in low-level (925–700 hPa) cloud over the tropical Indo-Pacific under warming, which can explain about 90% of λSWCL in CAMS-CSM. Static stability of planetary boundary layer in the pre-industrial simulation is a critical factor controlling the low-cloud response and λSWCL across the CMIP5 models and CAMS-CSM. Evidently, weak stability in CAMS-CSM favors lowcloud formation under warming due to increased low-level convergence and relative humidity, with the help of enhanced evaporation from the warming tropical Pacific. Consequently, cloud liquid water increases, amplifying cloud albedo, and eventually contributing to the unusually negative λSWCL and low ECS in CAMS-CSM. Moreover, the OHU may influence climate feedbacks and then the ECS by modulating regional sea surface temperature responses.