We develop a recycling model using C-13/C-12 mass balance for net growth/loss of the sedimentary organic carbon (C-org) reservoir, and apply it to the Neogene bulk marine carbonate delta(13)C record. The model allows for variations in photosynthetic fractionation factors, carbon cycling rates, and the isotopic composition of riverine carbon inputs to the oceans, The sign of the net flux term is controlled by the difference between fractional C-org burial and fractional C-org weathering, independent of any variations in carbon cycling rate, These terms are in turn estimated from the carbon isotope mass balance of newly deposited and weathered sediments, respectively, The magnitude of the net flux is sensitive to the global carbon cycling (erosion/deposition) rate, which may be estimated by the use of the records of radiogenic isotopic variations (Nd, Sr) in paleoseawater. A key observation and input to the model is that photosynthetic carbon isotope fractionation by both marine algae and terrestrial plants has decreased during the Cenozoic, Incorporating time-dependent carbon isotope fractionation into the model shows that the sedimentary C-org reservoir has grown throughout most of the Neogene, even as marine delta(13)C values fell after 14 Ma, A similar result is obtained if the variation in the marine delta(13)C record is largely caused by changes in the carbon isotopic composition of river fluxes to the oceans, rather than changes in the organic/inorganic ratio of output to the burial sink. The growth of the sedimentary organic carbon reservoir requires that the Neogene sedimentary carbon cycle was a net source of O-2 and a net sink of CO2 to the atmosphere, at least until the Plio-Pleistocene. As a consequence, Neogene CO2 consumption by silicate weathering cannot be balanced by oxidation of sedimentary C-org, placing a significant constraint on global carbon balance models. A related prediction of our model is that atmospheric O-2 levels rose during the Neogene.