In this paper, we propose using an University Entrance Exam (the Exame Nacional do Ensido Médio) as a proper test of success of artificial intelligence techniques, thus replacing the famous Turing Test. We argue that more importantly than measuring the ability of a system in replicating human competence, the so-called ENEM test can serve as a driver of development of new techniques. Additionally, we describe how we produced a machine-readable database of questions from previous exams, which can be used for comparing techniques for natural language processing, image processing and knowledge representation and reasoning. We then present some preliminaries to serve as baseline that are based on information retrieval techniques and Word2Vec. Experiments with previous exams show that in questions concerning Humanities and Languages these baseline methods perform in average slightly better than random guessing.