Cobalt, in the form of cobalt metal, salts, hard-metals and alloys, is widely used in a variety of industrial, medical and military applications. Chronic inhalation exposure to cobalt metal and cobalt sulfate have been shown to cause lung cancer in rats and mice, as well as systemic tumors in rats. Epidemiological studies have not found an association between exposure to cobalt metal or salts and increased cancer rates; however, there is evidence of carcinogenicity among workers exposed to hard metals where cobalt, tungsten and carbon are combined. The U.S. Environmental Protection Agency has derived a provisional inhalation unit risk (IUR) value of 0.009 µg/m3 for cobalt based on a chronic inhalation study of cobalt sulfate heptahydrate (a soluble salt) in rodents; however, a recent 2-year study of cobalt metal particulates affords the opportunity to derive IUR values specifically for cobalt metal. Based on benchmark dose modeling of lung tumors in the rodent bioassays, IUR values for cobalt metal ranged from 0.003 to 0.006 µg/m3, indicating that cobalt metal is a less potent lung carcinogen in rats than cobalt sulfate heptahydrate. Multistage and smoothing spline regression modeling gave comparable results, although the latter model predicted more nonlinearity in the low-dose region, which is not evident in the multistage model. In vivo and in vitro data, including mutation data from the recent bioassay, support that the carcinogenic mode of action (MOA) of cobalt metal in the lung is likely to involve oxidative stress, and thus, the exposure-response may be thresholded. Accordingly, low-dose linear extrapolation from the BMCL10 in high-dose animal studies likely over predicts the risk associated with occupational and environmental exposures. More work is needed to measure and model precursor events, which could support development of a reference concentration (RfC) protective of lung cancer.