New requirements for specification and forecast of the space environment and the availability of unprecedented amounts of real-time data are now driving the development of data assimilation schemes for the thermosphere and ionosphere. Such schemes require accurate knowledge of any biases affecting the models. Finding the biases is not trivial and requires significant effort. Here we present a first step in the validation of a coupled thermosphere ionosphere general circulation model in preparation for its inclusion in a data assimilation scheme. We present a comparison between the Mass Spectrometer Incoherent Scatter (MSIS) radar empirical model neutral temperatures and the Coupled Thermosphere Ionosphere Plasmasphere Electrodynamics (CTIPE) neutral temperature predictions for three solar cycle conditions (F10.7 = 70, 125, and 200), three geomagnetic activity conditions (Kp = 1, 3, and 7), and three seasons (equinox, summer, and winter). The CTIPE model was run for each case with constant inputs until a diurnally reproducible (“steady state”) global temperature pattern was obtained. MSIS predictions were generated for “perpetually constant” equivalent conditions. The temperature comparisons are performed on a 300 km altitude shell. We present global temperature averages, area-weighted root mean square differences, and zonally averaged temperature comparisons. CTIPE temperatures at 300 km altitude are lower than MSIS if Joule heating calculations do not include small-scale E field variability. This is the first global assessment of a general circulation model for the thermosphere over such a wide range of geomagnetic and solar conditions.