Résumé: A growth model from which the expected age of a fish can be estimated based on its length is a key component to most stock assessments. For the three tropical tuna species in the Indian Ocean – yellowfin (YFT; Thunnus albacares), bigeye (BET; T. obesus) and skipjack (SKJ; Katsuwonus pelamis) – information about growth has been very limited until recently, when data from a large-scale Indian Ocean tuna tagging program became available. In this paper, parametric growth models were fit to tag-recapture data for all three species using a maximum likelihood method that models the joint density of release and recapture lengths as a function of age by treating age at tagging as a random variable. The method allows for individual variability in growth by modelling the asymptotic length parameter as a random effect. Direct age and length data from otolith readings were also included in the analysis for YFT and BET. The results support two-stanza growth models for all three species; however, the growth patterns for YFT and BET differ from SKJ. YFT and BET exhibit a transition in growth between age 2 and 3, with faster growth in the second stanza than the first, whereas SKJ exhibit a transition in growth around age 1, with much faster growth in the first stanza than the second. Most likely, YFT and BET also experience a phase of rapid growth directly following hatching, but lack of data for fish less than 50 cm for these species precludes its estimation. Differences in growth between sexes were found for YFT and BET, with males growing to a larger size; information on sex was not available for SKJ.