The study used a hybrid approach to evaluate the performance of 20 global climate models (GCMs) from CMIP6 in reproducing 20 extreme temperature and precipitation indices over Pakistan. This study first analysed the future simulations of extremes and discarded the GCMs whose projections fell outside the 95% confidence interval. The remaining GCMs' performance was evaluated using the Kling Gupta Efficiency‐based criteria. The changes in extreme climatic events in Pakistan were projected using the multi‐model ensemble (MME) median of the selected GCMs for four shared socioeconomic pathways (SSPs) in two future periods, 2020–2059 and 2060–2099. Four GCMs' showed inconsistency in projecting future climatic extremes and were discarded initially. The past performance of the remaining GCMs revealed EC‐Earth3‐Veg‐LR, GFDL‐ESM4, MRI‐ESM2‐0 and NOR‐ESM2‐MM to be effective in reproducing extreme indices for the historical period. The MME median of the selected GCMs showed a gradual increase in most of the temperature and precipitation extreme indices in the future periods for all SSPs across the country. Specifically, it showed a higher increase in daily maximum temperature (TXx) by 4.5–5°C, daily minimum temperature (TNn) by more than 4.5°C, consecutive days with a temperature higher than the 95th percentile (WSDI) by >160 days, one‐day maximum rainfall by 9–15 mm and days with precipitation above the 95th percentile by >50 mm in northern high‐elevated areas during 2060–2099 for SSP585. Similarly, a higher increase in TXx by 4.5–5°C, TNn by >4.5°C, WSDI by 140 days and tropical nights by 40–60 days was also found in the western arid region during 2060–2099 for SSP585. The study highlights that northern high‐elevated and western arid regions are at a higher risk of extreme temperatures and precipitation due to climate change.