Carbon monoxide (CO) is an endogenous signaling molecule with demonstrated ability to modulate immune responses and to engage key components of the circadian clock. Further, CO has been pharmacologically validated for its therapeutic benefits in animal models of various pathological conditions. For the development of CO-based therapeutics, new delivery forms are needed to address the inherent limitations of using inhaled CO for therapeutic applications. Along this line, there have been metal– and borane–carbonyl complexes reported as CO-release molecules (CORMs) for various studies. CORM-A1 is among the four most widely used CORMs in examining CO biology. Such studies are predicated on the assumptions that CORM-A1 (1) releases CO efficiently and reproducibly under commonly used experimental conditions and (2) does not have meaningful CO-independent activities. In this study, we demonstrate the important redox properties of CORM-A1 leading to the reduction of bio-relevant molecules such as NAD⁺ and NADP⁺ under near-physiological conditions; such reduction reciprocally facilitates CO release from CORM-A1. We further demonstrate that CO-release yield and rate from CORM-A1 are highly dependent on various factors such as the medium used, buffer concentrations, and redox environment; these factors seem to be so idiosyncratic that we were unable to find a uniform mechanistic explanation. Under standard experimental conditions, CO release yields were found to be low and highly variable (0.5–15%) in the initial 15 min unless in the presence of certain reagents, e.g. NAD⁺ or high concentrations of buffer. The significant chemical reactivity of CORM-A1 and the highly variable nature of CO release under near-physiological conditions suggest the need for much more consideration of appropriate controls, if available, and caution in using CORM-A1 as a CO surrogate in biological studies.