We introduce and evaluate a novel network-based approach for determining individual credit of coauthors in multi-authored papers. In the proposed model, coauthorship is conceptualized as a directed, weighted network, where authors transfer coauthorship credits among one another. We validate the model by fitting it to empirical data about authorship credits from economics, marketing, psychology, chemistry, and biomedicine. Also, we show that our model outperforms prior alternatives such as fractional, geometric, arithmetic, and harmonic counting in generating coauthorship credit allocations that approximate the empirical data. The results from the empirical evaluation as well as the model’s capability to be adapted to domains with different norms for how to order authors per paper make the proposed model a robust and flexible framework for studying substantive questions about coauthorship across domains.