Commercially manufactured meters that measure the flow of a process gas are often calibrated with a known flow of a surrogate gas. This requires an accurate model of the flow meter and accurate values of the relevant thermophysical properties for both gases. In particular, calibrating a “laminar” flow meter near ambient temperature and pressure requires that the ratio (process gas viscosity)/(surrogate gas viscosity) be known to approximately 0.1%. With this motivation, we critically reviewed measurements of viscosity conducted with 18 instruments near 25 °C and zero density for 11 gases: He, Ne, Ar, Kr, Xe, H2, N2, CH4, C2H6, C3H8, and SF6. For these gases and this single state, we determined viscosity ratios with relative standard uncertainties ranging from 2.7 × 10−4 to 3.6 × 10−4 at a 68% confidence level. Anchoring the ratios to the value (19.8253 ± 0.0002) × 10−6 Pa s for the viscosity of helium calculated ab initio at 25 °C and zero density yields recommended values for the other ten gases and establishes a scale for gas viscosities that is more accurate than most of the reported values. To facilitate the extension of this scale, we recommend that researchers who calibrate gas viscometers (1) use helium as a calibration gas when possible, (2) report the values of all calibration data, and (3) report the uncertainties of their measured viscosity ratios. Similarly, we recommend that data archives capture this relevant calibration information.