<p>This study introduces the complex network theory for assessing the performance of General Circulation Models (GCMs) for studying meteorological droughts. The shortest path length, and particularly node efficiency, is used as a network-based measure for performance assessment. The performance of 53 CMIP6 GCMs for simulating meteorological droughts in India is assessed by comparing the node efficiency of the Standardized Precipitation Index (SPI) values obtained from GCM rainfall outputs with that of the observed (gridded and interpolated) data. Data at 1° x 1° spatial resolution across India (a total of 288 grids) are considered, and the GCMs are ranked separately for each grid based on the difference in node efficiency. A Group Decision Making (GDM) approach is employed to integrate the GCM performance across all grids. Four different timescales of SPI values (1, 3, 6, and 12 months) are also considered, and the final ranking is determined using a Comprehensive Rating Metric (RM) that consolidates results across the four different timescales. The results reveal significant variability in model performance across the four timescales. For instance, the models NorESM2-MM, CESM2-FV2, and KACE-1-0-G consistently emerge as top performers across the four timescales. Conversely, some models, such as MPI-ESM1-2-LR, FGOALS-g3, and CanESM5, exhibit less consistent rankings for the different timescales. A comparison of ranking of GCMs based on raw rainfall and on SPI values reveals that some models, such as IPSL-CM5A2-INCA, NorESM2-LM, and CMCC-ESM2, perform consistently well for both cases, while some others, such as FGOALS-g3, CanESM5, and GISS-E2-2-G, exhibit poor rankings for both rainfall and SPI. The outcomes from this study indicate the utility of complex network theory for assessing the performance of GCMs for drought-related studies based on drought indices and can supplement and complement studies that assess the performance of GCMs based only on raw rainfall data.</p>