 This research examined the benefits of combining four different solar radiation models into a single system. Two satellite image-based models and two data-driven models were compared with an all-sky imager, ASI, based model. The results showed that the best combination of models was the general blend model and random forest, RF, blend model. These models provided improved accuracy over the individual models, with the greatest improvements seen at lead times of 15 minutes or less. Additionally, the study found that the use of satellite imagery-based models could be substituted for the more costly ASI-based models without sacrificing accuracy. This article was authored by Miguel Lopez-Questa, Ricardo Alamor, Ines Maria Galvan-Leon and others.