Evaluation of ENSO in CMIP5 and CMIP6 models and its significance in the rainfall in Northeast Thailand

El Niño Southern Oscillation (ENSO) is a significant form of internal climate variability resulting from the interactions between the atmosphere and ocean in the tropical Pacific. It is a main driver of interannual rainfall variability in Northeast Thailand, where rainfed agriculture is one of the largest economic sectors. Therefore, it is essential to understand the ability of climate models to simulate the basic characteristics of ENSO phenomena and project its impacts in the region. We evaluated the ability of 12 climate models from the sixth phase of the Coupled Model Intercomparison Project (CMIP6) and their 12 predecessor models from the fifth phase (CMIP5) to simulate ENSO and its impact on rainfall over Northeast Thailand using thirteen performance metrics under seven criteria considering the simulation of observed indices, pattern, variability, peaks and seasonal phase locking of ENSO. The intensity and frequency of the ENSO events were projected for the near future (2021–2050). Although six CMIP5 and eight CMIP6 models performed well in simulating half of the ENSO evaluation metrics (performance score > 40%) such as ENSO variability, seasonal phase locking, location of ENSO variability, and dominant secondary peak in SST variability, we did not find very compelling evidence of improved ENSO simulation in CMIP6 compared to CMIP5 models. However, high rainfall in extreme La Niña events and low rainfall in extreme El Niño events with rainfall anomalies of 0.4 mm/day and − 0.3 mm/day, respectively were observed over Northeast Thailand due to the interaction of the easterly winds from the Pacific Ocean with the south-westerly flow from the Indian Ocean. The corresponding sea level pressure maps further confirmed the mechanisms associated with this phenomenon over the region. The ensemble averages of models from both phases predicted an increase in intensity (by 32–47%) and frequency (by 10–50%) in the near-future (2021–2050) with a slightly higher increase by CMIP6 models compared to CMIP5 models under medium and high emission scenarios.