Atmospheric carbon dioxide concentrations (pCO2) beyond ice core records have been reconstructed from δ11B derived from planktic foraminifera found in equatorial sediment cores. Here, I applied a carbon cycle model over the Plio-Pleistocene to evaluate the assumptions leading to these numbers. During glacials times, simulated atmospheric pCO2 was unequilibrated with pCO2 in the equatorial surface ocean by up to 35 ppm while the δ11B-based approaches assume unchanged (quasi)equilibrium between both. In the Pliocene, δ11B-based estimates of surface ocean pH are lower in the Pacific than in the Atlantic resulting in higher calculated pCO2. This offset in pH between ocean basins is not supported by models. To calculate pCO2 in surface waters out of the δ11B-based pH some assumptions on either total alkalinity or dissolved inorganic carbon are necessary. However, the assumed values of these under-constrained variables were according to my results partly inconsistent with chemically possible combinations within the marine carbonate system. The model results show glacial/interglacial variability in total alkalinity of the order of 100 μmol/kg, which is rarely applied to proxy reconstructions. Simulated atmospheric pCO2 is tightly (r2 > 0.9) related to equatorial surface-ocean pH, which can be used for consistency checks. Long-term trends in volcanic CO2 outgassing and the strength of the continental weathering fluxes are still unconstrained, allowing for a wide range of possible atmospheric pCO2 across the Plio-Pleistocene. Nevertheless, this carbon cycle analysis suggests that reported atmospheric pCO2 above 500 ppm in the Pliocene might, for various reasons, need to be revised to smaller numbers.