Alzheimer’s disease (AD) is the leading cause of dementia, and its early detection is key for accurate diagnosis and treatment. Plasma phosphorylated tau 217 (pTau217) has emerged as a promising non-invasive biomarker. This study aimed to establish a population-specific cut-off for plasma pTau217 levels in Argentina using the GeNED.ar cohort (Genetics and Neuroimaging of Aging and Dementia in Argentina). We analyzed data from 137 individuals: 83 cognitively unimpaired (CU), 34 with mild cognitive impairment (MCI), and 20 with AD. Finite Mixture Modelling was used to define thresholds without relying on gold-standard methods such as PET or CSF biomarkers. We estimated both a single standard cut-off and a two-threshold approach at probability levels of 0.8, 0.9, and 0.95. The latter defines a lower threshold to identify a high-sensitivity zone (likely negative) and an upper threshold for a high-specificity zone (likely positive), with an intermediate “gray zone” requiring further testing. The two-threshold method at the 0.9 probability level showed the best performance, with lower and upper thresholds of 0.47 and 0.83, respectively. These thresholds are comparable to those reported in other populations. The results also support the clinical diagnoses established at the Memory Clinic. Our findings reinforce the utility of plasma pTau217 as a scalable and accessible diagnostic tool for AD.