In calculating solar radiation, climate models make many approximations to the known physics of radiative transfer. These simplifying parameterizations are made to reduce computational cost and enable climate modeling, but they obviously cause errors in solar heating that impact the simulated climate. Most of these radiative transfer errors have been identified individually in isolated examples, but here we quantify them in terms of net solar heating of the atmosphere and surface within a consistent framework on a scale relevant to the global climate. We build a benchmark capability around a solar heating code (Solar-J) that already includes some of the more accurate radiative transfer methods and add further improvements covering known errors. The error classes assessed here include: use of broad wavelength bins to integrate over fine spectral features; multiple-scattering approximations that alter the scattering phase function and optical depth for clouds, aerosols, and gases; uncertainty in ice-cloud optics; treatment of fractional cloud cover including cloud overlap; and constant ocean surface albedo. We geographically map the errors in terms of W m using a full climate re-creation for January 2015 from weather forecasting models. For many of the ten specific approximations calculated here, the mean errors are ~2 W m with even larger latitudinal biases and are likely to affect a model’s ability to match the current climate state. From this study, we are able to make priority recommendations for these errors, pointing out where codes can be simply updated and where more scientific development is needed.