Cyanide has several antidotes, with differing mechanisms of action and diverse toxicological, clinical, and risk-benefit profiles. The international medical community lacks consensus about the antidote or antidotes with the best risk-benefit ratio. Critical assessment of cyanide antidotes is needed to aid in therapeutic and administrative decisions that will improve care for victims of cyanide poisoning (particularly poisoning from enclosed-space fire-smoke inhalation), and enhance readiness for cyanide toxic terrorism and other mass-casualty incidents. This paper reviews preclinical and clinical data on available cyanide antidotes and considers the profiles of these antidotes relative to properties of a hypothetical ideal cyanide antidote. Each of the antidotes shows evidence of efficacy in animal studies and clinical experience. The data available to date do not suggest obvious differences in efficacy among antidotes, with the exception of a slower onset of action of sodium thiosulfate (administered alone) than of the other antidotes. The potential for serious toxicity limits or prevents the use of the Cyanide Antidote Kit, dicobalt edetate, and 4-dimethylaminophenol in prehospital empiric treatment of suspected cyanide poisoning. Hydroxocobalamin differs from these antidotes in that it has not been associated with clinically significant toxicity in antidotal doses. Hydroxocobalamin is an antidote that seems to have many of the characteristics of the ideal cyanide antidote: rapid onset of action, neutralizes cyanide without interfering with cellular oxygen use, tolerability and safety profiles conducive to prehospital use, safe for use with smoke-inhalation victims, not harmful when administered to non-poisoned patients, easy to administer.