Despite being a target for about one-third of approved drugs, G protein-coupled receptors (GPCRs) still represent a tremendous reservoir for therapeutic strategies against countless diseases. For example, several cardiovascular and central nervous systems conditions could benefit from clinical agents that activate the adenosine 1 receptor (A1R), however, the pursuit of A1R agonists for clinical use are usually impeded by both on- and off-target side effects. One of the possible strategies to overcome this issue is the development of positive allosteric modulators (PAMs) capable of selectively enhancing the effect of a specific receptor subtype and triggering functional selectivity (a phenomenon also referred to as bias). Intriguingly, besides enforcing the effect of agonists upon binding to an allosteric site, most of the A1R PAMs display intrinsic partial agonism and orthosteric competition with antagonists. To rationalize this behaviour, we simulated the binding of the prototypical PAMs PD81723 and VCP171, the antagonist 13B, and the bitopic agonist VCP746. We propose that a single PAM can bind several A1R sites rather than a unique allosteric pocket, reconciling the structure-activity relationship and the mutagenesis results.