To support building design in view of daylight quality, computational (simulation) tools can provide effective support. To perform detailed daylight analysis via simulation, appropriate sky models are needed. In the past, various sky luminance distribution models have been developed. Such models, however, require illuminance data for the relevant location. As measured external illuminance levels are not available for many locations, the more widely available irradiance measurements can be translated, using proper luminance efficacy functions, into illuminance values. The present paper compares five global luminous efficacy models based on a database of measured illuminance and irradiance data from Vienna, Austria. These models typically involve mathematical formulations with multiple coefficients, whose values are derived for a specific location. The results suggest that these models performed rather poorly, once tested against Vienna data. However, the models’ performance improved significantly, once the respective coefficients were modified (calibrated) using the Viennese database.