Calculations of air‐mass factors (AMFs) for ground‐based zenith‐sky UV‐visible spectrometers are now well developed in laboratories where stratospheric constituents are measured with this technique. An intercomparison between results from the different radiative transfer models used to calculate AMFs at twilight is presented here. The comparison was made for ozone AMFs at 510 nm and for NO2 AMFs at 440 nm. Vertical profiles were specified. Results are presented firstly for calculations in a pure Rayleigh atmosphere, then including background aerosols. Relative differences between calculated AMFs from different models cause relative errors in vertical columns of ozone and NO2 measured by zenith‐sky spectrometers. For commonly used averages over solar zenith angles, these relative errors are ±2.3% in the vertical column of ozone and ±1.1% in the vertical column of NO2. Refinements to the calculations, suggested by the intercomparison, should reduce these errors to ±1.0% for ozone and ±0.5% for NO2.