Objectives We examined the risk of mortality and cancer incidence with quantitative exposure to benzene-soluble fraction (BSF), benzo(a)pyrene (BaP), fluoride, and inhalable dust in two Australian prebake smelters. Methods A total of 4,316 male smelter workers were linked to mortality and cancer incidence registries and followed from 1983 through 2002 (mean follow-up: 15.9 years, maximum: 20 years). Internal comparisons using Poisson regression were undertaken based on quantitative exposure levels. Results Smoking-adjusted, monotonic relationships were observed between respiratory cancer and cumulative inhalable dust exposure (trend p = 0.1), cumulative fluoride exposure (p = 0.1), and cumulative BaP exposure (p = 0.2). The exposure–response trends were stronger when examined across the exposed categories (BaP p = 0.1; inhalable dust p = 0.04). A monotonic, but not statistically significant trend was observed between cumulative BaP exposure and stomach cancer (n = 14). Bladder cancer was not associated with BaP or BSF exposure. No other cancer and no mortality outcomes were associated with these smelter exposures. Conclusions The carcinogenicity of Söderberg smelter exposures is well established; in these prebake smelters we observed an association between smelter exposures and respiratory cancer, but not bladder cancer. The exploratory finding for stomach cancer needs confirmation. These results are preliminary due to the young cohort and short follow-up time.