This study evaluated the efficacy and safety of unengineered tumor-infiltrating lymphocytes (TILs) combined with pembrolizumab and either high (HD, Arm-1) or low (LD, Arm-2) doses of IL-2 in patients with metastatic melanoma (MM). Patients were lymphodepleted with cyclophosphamide and fludarabine, followed by TIL infusion and IL-2 (Arm-1: 720,000 IU/kg IV q 8 hrs up to 15 doses; Arm-2: 2 million IU SC for 14 days). Patients received pembrolizumab 200 mg IV starting 21 days post-TIL infusion, and every 3 weeks for up to 2 years. The primary endpoint was overall response rate (ORR) per RECIST 1.1. Blood samples were collected for longitudinal flow cytometry and cytokine analysis. In Arm-1 (n = 7), one patient had a partial response (PR) for 10 months, two had stable disease (SD), three had progressive disease (PD), and one was not evaluable (NE). In Arm-2 (n = 7), one patient had an ongoing PR for over 76 months, one had SD, and five had PD. The toxicity profiles were comparable; however, patients in Arm-2 had lower grade 3 febrile neutropenia (57% vs. 71%) and shorter hospitalization (median 16 days vs. 18 days). No correlation was observed between TIL phenotype and clinical response, although PR patients received high numbers of TIL with a high CD8+/CD4+ T cell ratio. IL-2 dose did not affect the frequency, phenotype, or proliferation of circulating T cell subsets, and anti-PD-1 did not boost T-cell proliferation. No significant differences were observed between IL-2 doses, suggesting low-dose IL-2 as an alternative to high-dose IL-2 after TIL administration.