Sea-ice thickness, though critical to our understanding of sea-ice variability, remains relatively understudied compared to surface sea-ice parameters in the Southern Ocean. To remedy this, we examine spatio-temporal variations in sea-ice thickness by analyzing historical simulations from 39 coupled climate models in CMIP6, comparing them with three sea-ice products, including satellite-derived observations and reanalyses. Furthermore, we compare seasonal trends in simulated sea ice thickness with trends in sea ice area. Our results indicate that CMIP6 models can replicate the mean seasonal cycle and spatial patterns of sea-ice thickness. During its maximum in February, these models align well with satellite-based observation products. However, during the annual minima, CMIP6 models show significant agreement with the reanalysis products. Certain models exhibit unrealistic historical mean states compared to the sea-ice products resulting in significant inter-model spread. CMIP6 models can simulate sea-ice area more accurately than the sea-ice thickness. They also simulate a positive relationship between the two parameters in September such that models with greater area tend to exhibit thicker ice. In contrast, there is a negative relationship in February when greater area is associated with lower thickness since only the thicker ice survives the summer melt. Moreover, our study highlights significant positive trends in sea-ice thickness observed during the cooler seasons, which are nearly absent in the warmer seasons where positive trends are predominantly observed in sea-ice area. While CMIP6 models perform well in simulating sea-ice area and its relationship with thickness, accuracy in the latter remains a challenge. This study, therefore, highlights the need for improved representation of Antarctic sea-ice processes in models for accurate projections of thickness and related volume changes.