Present knowledge about the fate of persistent and mobile (PM) substances in drinking water treatment is limited. Hence, this study assesses the potential of fixed-bed granular activated carbon (GAC) filters to fill the treatment gap for PM substances and the elimination predictability from lab-scale experiments. Two parallel pilot filters (GAC bed height 2 m, diameter 15 cm) with different GAC were operated for 1.5 years (ca. 47,000 BV throughput) alongside rapid small-scale column tests (RSSCT) designed based on the proportional diffusivity (PD) and the constant diffusivity (CD) approaches. Background dissolved organic matter (DOM) and a set of 17 target substances were investigated, among them 2-acrylamido-2-methylpropane sulfonate (AAMPS), adamantan-1-amine (ATA), melamine (MEL) and trifluoromethanesulfonic acid (TFMSA). Nine substances were predominantly present in the drinking water used as pilot filter influent (frequencies of detection above 80 %, median concentrations 0.003–1.868 μg/L) and their breakthrough behaviors could be observed: TFMSA was not retained at all, four substances including AAMPS and ATA reached complete breakthrough below 20,000 BV, three compounds were partially retained until the end of operation and oxypurinol was retained completely. The comparable PM candidate and DOM removal performances of both GAC aligns with their very similar surface characteristics and elemental compositions. The agreement of results between RSSCT with the pilot-scale filters were substance specific and no superior RSSCT design could be identified. However, CD-RSSCT provide a conservative removal prediction for most studied compounds. MEL adsorption was significantly underestimated by both RSSCT designs. Using the criterion of a carbon usage rate (with respect to 50 % breakthrough) below 25 mg_GAC/L_water for an economic retention by fixed-bed GAC filters, five (out of nine) substances met the requirement.