Abstract: Techniques are described for generating a programming score for a user based on source code submissions for a programming task set. In an embodiment, a programming system generates, at random, a programming task set for submission by a user. The programming system receives source codes submitted for the programming task set in a programming session with a client computer system of a user. The programming system executes the source codes to evaluate their accuracy. Based on the execution, the programming system determines that some source codes are evaluated to be accurate thereby solving corresponding programming tasks from the programming task set. Based on such an execution, programming system determines the programming score for the particular user for the programming task set.