The Cross-track Infrared Sounder (CrIS) on board the Suomi National Polar-Orbiting Partnership (S-NPP) satellite is a hyperspectral Fourier Transform Spectrometer. In this study, biases of the 399 channels used in numerical weather prediction applications are characterized based on the differences between CrIS observations and model simulations in clear-sky conditions over ocean. The Visible Infrared Imaging Radiometer Suite (VIIRS) cloud mask is used for selecting CrIS clear-sky data. The global mean biases are within ±1K for most channels (more than 390). Biases for the channels not affected by trace gases other than water vapor in long-wave infrared (LWIR), middle-wave infrared (MWIR) and short-wave infrared (SWIR) bands vary from about −0.5 to 0.3K, −0.2 to 0.8K, and −0.1 to 0.9K, respectively. The scan-angle variations of biases are less than ±0.4K for all channels. The MWIR sounding channels have larger biases in middle and high latitudes than the tropics, which might be associated with latitudinal differences of water vapor variability. The SWIR CO2 channels are affected by nonlocal thermodynamic equilibrium (NLTE) in the upper stratosphere and the window channels could be affected by the sun glint effect, both of which are considered in model simulations. Biases of the NLTE affected channels during daytime are found to be 0.5–1K higher than nighttime. The SWIR window channels have colder biases at Field of Regards (FORs) 6–15 than the other scan positions due to the sun glint effect.