This article addresses the question of whether race is a biological category and whether it is permissible to use it in biomedicine. I suggest that instrumentalism, a view that race is a problem-solving tool rather than a concept with an objective referent in nature, may be more consistent with the available scientific evidence. I argue that, to be morally permissible, the instrumentalist use of race in research and medicine requires stringent guidelines. I then provide four normative rules to guide race research in the biomedical sciences. The paper gathers evidence from philosophy of science, genomics, legal history, and normative ethics in order to ground the biomedical use of race in a converging ethical and epistemic framework.