A forthcoming version of the Energy Exascale Earth System Model (E3SM) will simulate global CO2 transport and interaction with surface biogenic and anthropogenic CO2 fluxes prognostically. Thorough comparison of the resulting CO2 fields against observations is essential to learning more about global biogeochemical dynamics and refining model components. However, currently available benchmarking and analysis tools for E3SM and other Coupled Model Intercomparison Project (CMIP) models do not provide comprehensive diagnostics for atmospheric CO2 utilizing multiple measurement platforms—such as surface-, satellite-, and aircraft-based observations. To facilitate such evaluations, we developed the open-source, Python-based Greenhouse Gas Diagnostics for Earth System Simulations (GDESS) tool, which provides automated parsing of observational datasets and model simulations to be in a consistent structure, computation of statistical metrics, and generation of diagnostic visualizations. GDESS enables scientists to evaluate a model’s ability to reproduce observed temporal and spatial variations of atmospheric CO2—e.g., multidecadal trends, seasonal cycles, interannual variability, and horizontal and vertical gradients. We intend for the set of diagnostics in GDESS to form a common framework for comparing simulated and observed concentrations of atmospheric CO2 and other greenhouse gases across Earth system models. These automated diagnostics will support more rapid evaluation and improvement of model-simulated global CO2 sources and sinks associated with land and ocean ecosystem processes.