Elasmobranchs make up a significant part of bycatch in pelagic longline fisheries, whose induced mortality can be a major threat to endangered species. It is therefore crucial to understand the drivers of at-vessel mortality (AVM) for this fishing gear in order to enhance post-release survival. To this end, we analysed scientific data collected during monitored longline fishing experiments conducted in French Polynesia to (i) estimate AVM for each species based on bootstrapped samples, and (ii) to assess AVM drivers using multivariate logistic regression models for the blue shark (Prionace glauca) and oceanic whitetip shark (Carcharhinus longimanus). We found that AVM varies widely between species. Oceanic whitetip sharks are more likely to die when caught in waters outside their comfort temperature range, and their odds of survival increase with body length. For the blue shark, the only driver related to AVM is hooking duration. These results indicate that to reduce AVM of these two species, the vertical distribution of hooks and soak duration should be considered as mitigation measures related to pelagic longlining.