In a classification model, the output is discrete, a set of categories. e.g. Does the person in this photo has curly or straight hair?
In a regression model, the output is continuous, a numeric quantity. e.g. What is the estimated salary of a person with these characteristics?