A chest radiograph is a chest projection radiograph that has been used to diagnose the disorders that affect the chest, its contents and structures in the vicinity. The chest X-ray of a pneumonia affected COVID-19 patient differs from a healthy person's chest X-ray. Differentiating between them is difficult for the untrained human eye. But deep learning networks can learn to distinguish these distinctions. This paper analyses the performance of seven different models: Xception, VGG-16, ResNet-101-V2, ResNet-50-V2, MobileNet-V2, DenseNet-121 and Inception-ResNet-V2, when differentiating between COVID-19 and normal chest X-rays. The experiment's results on the COVID-19 chest X-ray dataset inferred that the Xception model performed the best. Inception-ResNet-V2 worked fine after Xception.