As a result of both urbanization and climate change, the thermal environment of cities is deteriorating steadily. Thus, urban planner and city officials are increasingly under the pressure maintain livable environments in cities. The quickest and most economic means to assessing the performance of various urban heat mitigation strategies is via numerical modeling. Owing to these advantages, the approach gained popularity over the past decade, as well as resulted in the proliferation of microclimate models. The driving parameter of outdoor human thermal comfort is radiation, which is accounted for via mean radiant temperature (Tmrt). While this parameter is at the center of most outdoor thermal comfort indices, it is also rather challenging to measure and calculate. Along with the recent emergence of microclimate models, the ways of calculating outdoor Tmrt also multiplied. However, beyond individual model validations, very few studies have attempted to compare the performance of different models. Thus, the aim of this paper is to assess the performance of three microclimate models in estimating Tmrt in the urban environment. The reviewed models are VTUF-3D, the Grasshopper add-ons of Rhinoceros 3D and the latest ENVI-met version utilizing the Indexed View Sphere (IVS) algorithm. The adopted spatial resolution in each model were not identical. In the case of ENVI-met and VTUF-3D, 3 and 5 meter resolutions were used, respectively. The Grasshopper add-on was set to calculate surface temperatures at 6 m resolution and view factors (along with Tmrt) for a denser, 3-meter spatial grid. The model assessment utilized data from a 26-hours-long integral radiation measurement, conducted in Szeged (Hungary). The observations were undertaken along the four bounding facades of Bartók Square during a clear summer day in 2016. According to the preliminary results, ENVI-met performs the best when the observation sites are exposed to direct solar radiation. However, it underestimates Tmrt values by up to 10 ºC when the sites are in shade and during night. In general, both VTUF-3D and the Grasshopper add-on underestimate Tmrt values considerably during sunlit conditions (exceeding 20 ºC at times), but reproduce radiative conditions well at night (with errors remaining below 4 ºC). One reason for the above errors is the way Tmrt is calculate by these models. While Tmrt calculations differ in each model, at the core of their adopted approaches is a Tmrt estimation developed for indoor environments---which was later adopted to outdoors. One of the lingering simplification is the assumption of a seated posture and/or the conceptualization of the human body as a sphere. The other transferred simplification is the assumption of small temperature differences between the surrounding surfaces. These assumptions not only diminish the impact of horizontal fluxes, but also depict a less diverse radiative environment.