Hyperspectral infrared sounding contains information about clouds, which plays an important role in modulating Earth's climate. However, there is a great deal of uncertainty in modeling the radiative effect of clouds due to its complex dependence on various parameters. Therefore, cloudy scenarios are often neglected in retrievals of infrared spectral measurements and in data assimilation. One‐dimensional radiative transfer (RT) models have a limited capability to represent the cloud three‐dimensional multilayer structure. This issue is typically resolved by using a multiple independent column approach, which is computationally demanding. Therefore, it is necessary to find a balance between computational speed and accuracy for infrared RT all‐sky radiance simulations. In this study, we utilize the Community Radiative Transfer Model with four different cloud overlap schemes and compare against observations made by the Atmospheric Infrared Sounder (AIRS) using a statistical metric called the first Wasserstein distance. Our results show that the average cloud overlap scheme performs the best and successfully predicts the overall probability distribution of brightness temperature over nonfrozen oceans for a wide range of wavelengths. The mean absolute differences are less than 0.7 K for 846 selected AIRS channels between 790 cm−1 and 1231 cm−1. Plain Language Summary: Clouds have a major impact on Earth's climate. However, modeling the three‐dimensional effects of clouds is complicated and computationally intensive, necessitating approximations with respect to their vertical structure. We utilize a commonly used RT model and evaluate a variety of cloud overlap schemes against observations made by the Atmospheric Infrared Sounder. This work introduces two novelties to the evaluation. First, we perform comparisons for hundreds of wavelengths to exploit the hyperspectral nature of the measurements. Second, we utilize the first Wasserstein distance metric rather than the traditionally used Pearson correlation. We demonstrate that the former is a much better approach to compare probability distributions. The best performing overlap scheme obtains a very good match with measurements. Key Points: Hyperspectral simulations, for more than 80,000 scenarios, using different cloud overlap schemes are compared with Atmospheric Infrared Sounder (AIRS) observationsAverage overlap scheme provides best results compared with AIRS observationsFirst Wasserstein distance is a more suitable metric than the commonly used Pearson correlation for comparing probability distributions [ABSTRACT FROM AUTHOR]