Recent research demonstrated the existence of a combination mode (C-mode) originating from the atmospheric nonlinear interaction between the El Niño-Southern Oscillation (ENSO) and the Pacific warm pool annual cycle. In this paper, we show that the majority of coupled climate models in the Coupled Model Intercomparison Project Phase 5 (CMIP5) are able to reproduce the observed spatial pattern of the C-mode in terms of surface wind anomalies reasonably well, and about half of the coupled models are able to reproduce spectral power at the combination tone periodicities of about 10 and/or 15 months. Compared to the CMIP5 historical simulations, the CMIP5 Atmospheric Model Intercomparison Project (AMIP) simulations can generally exhibit a more realistic simulation of the C-mode due to prescribed lower boundary forcing. Overall, the multi-model ensemble average of the CMIP5 models tends to capture the C-mode better than the individual models. Furthermore, the models with better performance in simulating the ENSO mode tend to also exhibit a more realistic C-mode with respect to its spatial pattern and amplitude, in both the CMIP5 historical and AMIP simulations. This study shows that the CMIP5 models are able to simulate the proposed combination mode mechanism to some degree, resulting from their reasonable performance in representing the ENSO mode. It is suggested that the main ENSO periods in the current climate models needs to be further improved for making the C-mode better.