The 15-minute city model relies on accurate spatial data to assess proximity between residents and essential urban services. A multiscale, transversal, and replicable decision-support model cannot depend on manual surveys alone and must necessarily rely on web-based geographic databases. This paper investigates the reliability of two widely used Open Data sources—OpenStreetMap (OSM) and Google Maps (GM)—in supporting such modelling efforts. Through a comparative analysis of four European urban contexts (Versilia, Nice, Wien, and Gothenburg), the study evaluates data consistency in terms of volume, spatial distribution, and classification structure, across both macro (urban) and micro (neighbourhood) scales. Findings reveal substantial differences between the two datasets, underscoring how the choice of dataset—and how it is used—can significantly affect 15-minute city modelling outcomes, potentially leading to misleading analyses or decisions. Specifically, OSM tends to return a larger volume of data, including private and non-accessible amenities, while GM provides fewer but more curated features, though constrained by download limits. Additionally, semantic inconsistencies in place categorisation, especially for multifunctional or culturally specific amenities, further restrict cross-platform interoperability or reliable comparison. The study concludes that a context-sensitive evaluation of datasets is essential. This should be supported by scale-aware analyses, empirical validation through ground-truthing, and the development of data cleansing protocols or adaptive frameworks to strengthen decision-support tools and ensure the reliability of data used in urban planning processes.