The Pacific–Japan (PJ) pattern is the dominant teleconnection over the western North Pacific in boreal summer and exerts substantial control on East Asian temperature, precipitation, and marine heat extremes. Despite its importance, the PJ pattern is simulated with large inter-model spread in current climate models, and a systematic, process-based framework for its evaluation has been lacking. Here, we propose the PJ pattern as a process-based evaluation metric for East Asian summer weather and climate simulations. The framework decomposes the PJ into four physically meaningful components – horizontal circulation structure, physical source, wave properties, and weather impacts – that together form a coherent chain of causality from tropical forcing to regional climate response. We apply the framework to 23 CMIP6 historical simulations using a common basis function method that enables consistent inter-model comparison against ERA5. The CMIP6 models generally reproduce the spatial structure of the PJ with high pattern correlations (> 0.8), but exhibit a remarkably wide spread in amplitude (25–170 % of observations). This spread is explained by a coherent physical chain: convective and moisture variability over the subtropical western Pacific drives the wave source, which is transmitted to mid-latitudes via the meridional wave activity flux, and ultimately translates into downstream temperature, precipitation, and SST responses over East Asia. The proposed framework, summarized in a single hexagonal diagram, is reproducible and modular, and offers an efficient, physically interpretable diagnostic for assessing East Asian climate simulation capability in current and future model generations.