Five physical models of the ionosphere were compared with each other and with data obtained at the Millstone Hill Observatory. Two of the models were self-consistent ionosphere-thermosphere models, while for the other ionospheric models the thermospheric parameters were provided by empirical inputs. The comparisons were restricted to midlatitudes and low geomagnetic activity, but four geophysical cases were considered that covered both the summer and winter solstices at solar maximum and minimum. The original motivation of the study was to determine why several physical models consistently underestimated the F region peak electron density, by up to a factor of 2, in the midlatitude, daytime ionosphere at solar maximum. This problem was resolved, but the resolution did not identify a lack of physics in any of the models. Instead, various chemical reaction rates, photoionization processes, and diffusion coefficients had to be adjusted, with the main one being the adoption of the Burnside factor of 1.7 for the diffusion coefficients. The subsequent comparisons of the models and data were for standard simulations in which uncertain inputs or processes were not adjusted to get better agreement with the data. For these comparisons, the five models displayed diurnal variations that, in general, agreed with the measurements. However, each one of the five models exhibited a clear deficiency in at least one of the four geophysical cases that was not common to the other models. Therefore, contrary to expectations, the coupled ionosphere-thermosphere models were not found to be superior to the uncoupled ionospheric models for the cases considered. The spread in NmF2 calculated by the five models was typically less than a factor of 2 during the day but was as large as a factor of 10 at certain local times during the night. The latter problem was traced to insufficient nocturnal maintenance processes in two of the uncoupled ionospheric models. The general findings of this study have important implications for the National Space Weather Program. ¿ 1998 American Geophysical Union