Conditional instability and the buoyancy of plumes drives moist convection but have a variety of representations in model convective schemes. Analyses of vertical structure information from Atmospheric Radiation Measurement (ARM) sites and reanalysis products (ERA5 and TRMM3b42) using diagnostics of the environment relevant to plume buoyancy can help constrain climate models. Previous work (Emmenegger et al. 2022) has shown a cohort of CMIP6 models to represent moist convective processes more accurately than their CMIP5 counterparts. However, certain biases in convective onset pervasive among generations of CMIP modeling efforts—weak precipitation rates that occur too frequently—remain. We diagnose these biases by applying a plume model with different mixing assumptions to vertical profiles of temperature and humidity in a cohort of eleven CMIP6 models. The fast convective adjustment timescale is used to infer model entrainment from the vertical thermodynamic structures. Most models capture qualitative aspects of equivalent potential temperature vertical structure, including a substantial decrease in height in the lower free troposphere associated with the entrainment of subsaturated air. We define a “pseudo-entrainment” diagnostic that combines a lower-free-tropospheric lapse rate and subsaturation similar to what entrainment would produce under certain assumptions, capturing the trade-off between larger lapse rates (entrainment of dry air) and small subsaturation (permits positive buoyancy despite high entrainment). This pseudo-entrainment diagnostic is also a reasonable predictor of the critical value of integrated buoyancy for precipitation onset. Models with poor temperature lapse rates (those using variants of the Tiedtke Scheme) or low entrainment runs of CAM5, and models with low subsaturation, such as NASA-GISS, lie outside the observational range in this diagnostic.