Anthropogenic contamination of water resources poses a critical global challenge for designing efficient adsorption-based water treatment processes. This work provides a systematic adsorption study of six graphene-based sorbents with different C/O ratios and surface chemistries for the removal of four representative cationic and anionic dyes and two broad-spectrum antibiotics (amoxicillin and norfloxacin) from aqueous solutions, including their mixtures in model wastewater. Overoxidized graphene oxide demonstrates high removal efficiencies for cationic dyes, with maximum sorption capacities of 155 mg/g for methylene blue and 175.5 mg/g for brilliant green. In contrast, under-oxidized graphene oxide shows even higher capacities for several dyes, reaching 369, 444 and 488 mg/g for methylene blue, brilliant green and Congo red, respectively. Thermally reduced graphene oxide combines partial reduction of carbon framework with residual oxygen functionalities and achieves sorption capacities of 153 mg/g for methylene blue and 209 mg/g for methyl orange. All graphene-based sorbents also effectively remove antibiotics, with maximum sorption capacities in the range of 322–487 mg/g. Adsorption kinetics and equilibrium are described using pseudo-second-order and nonlinear Langmuir and Sips models, enabling quantitative comparison of sorbent performance. The effectiveness of the graphene-based sorbents is further validated under near-real conditions in river water purification experiments, where dyes were present at extremely low concentrations (1 ppm). These results demonstrate that controlled oxidation and reduction of graphene oxide provide an effective route for tailoring sorbent selectivity and adsorption performance in wastewater treatment applications.