A comparative analysis of East Asian summer monsoon (EASM) precipitation is performed to reveal the drivers and mechanisms controlling the similarities of the mid-Pliocene EASM precipitation changes compared to the corresponding pre-industrial (PI) experiments derived from atmosphere-only (i.e. AGCM) and fully coupled (i.e. CGCM) simulations, as well as the large simulated differences in the mid-Pliocene EASM precipitation between the two simulations. The area-averaged precipitation over the EASM domain is enhanced in the mid-Pliocene compared to the corresponding PI experiments performed by both the AGCM (LMDZ5A) and the CGCM (IPSL-CM5A). Moisture budget analysis reveals that it is the surface warming over East Asia that drives the area-averaged EASM precipitation increase in the mid-Pliocene in both simulations. The surface warming increases the atmospheric moisture content, as revealed by an increase in the thermodynamic component of vertical moisture advection, resulting in enhanced mid-Pliocene EASM precipitation compared to PI in both simulations. Moist static energy diagnosis identifies the combined effect of enhanced zonal thermal contrast and column-integrated meridional stationary eddy velocity $\overline \upsilon ^*$and its convergence as the physical mechanisms that sustain the enhancement of mid-Pliocene EASM precipitation in both simulations compared to the PI experiments. This takes place through a strengthening of the EASM circulation and moisture transport into the EASM domain associated with an increase in local moisture convergence in the mid-Pliocene in both simulations. Moisture budget analysis also reveals that the larger area-averaged mid-Pliocene EASM precipitation increase in the CGCM compared to its AGCM component is mainly caused by the dynamical component contributing more to the vertical moisture advection in the CGCM (i.e. IPSL-CM5A) compared to its AGCM (LMDZ5). The large simulated differences in the spatial pattern of the mid-Pliocene EASM precipitation between the two simulations result from the combined effect of enhanced meridional thermal contrast over the EASM domain and increased convergence over South China in the CGCM simulation compared to the AGCM simulation.