Abstract This study examines the climatology and dynamics of atmospheric blocking, and the general circulation features that influence blocks in GFDL’s atmosphere-only (AM4) and coupled atmosphere–ocean (CM4) comprehensive models. We compare AM4 and CM4 with reanalysis, focusing on winter in the Northern Hemisphere. Both models generate the correct blocking climatology and planetary-scale signatures of the stationary wave. However, at regional scales some biases exist. In the eastern Pacific and over western North America, both models generate excessive blocking frequency and too strong of a stationary wave. In the Atlantic, the models generate too little blocking and a weakened stationary wave. A block-centered compositing analysis of block-onset dynamics reveals that the models 1) produce realistic patterns of high-frequency (1–6-day) eddy forcing and 2) capture the notable differences in the 500-hPa geopotential height field between Pacific and Atlantic blocking. However, the models fail to reproduce stronger wave activity flux convergence in the Atlantic compared to the Pacific. Overall, biases in the blocking climatology in terms of location, frequency, duration, and area are quite similar between AM4 and CM4 despite the models having large differences in sea surface temperatures and climatological zonal circulation. This could suggest that other factors could be more dominant in generating blocking biases for these GCMs. Significance Statement Atmospheric blocks are persistent high pressure systems that can lead to hazardous weather. Historically, climate models have had trouble capturing blocks, but recent changes in the models might lead to improvements. As such, the work herein investigates the spatial distribution, prevalence, duration, size, and dynamics of wintertime blocking in recent NOAA climate models. Overall, these models capture the long-term-average spatial pattern of blocking, and properly reproduce key dynamical features. However, the models produce too much blocking in the western United States, and too little over the northern Atlantic Ocean and Europe. These blocking biases are consistent with atmospheric stationary waves biases, but not jet stream bias. This downplays the role of jet biases in the models being responsible for blocking biases.