OBJECTIVES: Develop a Markov model to determine the cost-effectiveness of nivolumab, ipilimumab, and nivolumab-ipilimumab combination as first-line therapy in metastatic melanoma accounting for differential effectiveness of PD-L1 positive and negative states.  METHODS: A three-state Markov model (PD-L1 positive stable disease, PD-L1 negative stable disease, and progression/death) was developed using a US societal perspective with a lifetime time horizon of 14.5 years. Transition probabilities were progression-free survival data from the CheckMate-067 trial. Costs were expressed in 2015 US dollars and were determined using national sources. Adverse event (AE) management was determined using immune-related AE (irAE) data from CheckMate-067, irAE management guides for nivolumab and ipilimumab, and treatment guidelines. Utilities were obtained from published literature, using melanoma-specific studies when available, and were weighted based on incidence and duration of health states. Probabilistic sensitivity analyses were conducted. RESULTS: In the base case, ipilimumab monotherapy was dominated by nivolumab monotherapy while nivolumab-ipilimumab combination therapy had an ICER of $883,994/QALY compared to nivolumab. The cost of ipilimumab would have to decrease to <$4,500 per dose for combination therapy to be cost-effective, and to <$2,720 (a 90% reduction) to be more cost-effective than nivolumab alone.    CONCLUSIONS: At a willingness-to-pay threshold of $100,000/QALY, and using progression-free survival and current drug prices, nivolumab-ipilimumab combination therapy would not be cost-effective compared to nivolumab monotherapy, the most cost-effective option.