The oxygen isotope ratio (δ 18 O) of inorganic and biogenic carbonates is widely used to reconstruct past environments. However, the oxygen isotope exchange between CaCO 3 and H 2 O rarely reaches equilibrium and kinetic isotope effects (KIE) commonly complicate paleoclimate reconstructions. We present a comprehensive model of kinetic and equilibrium oxygen isotope fractionation between CaCO 3 and water (α c/w ) that accounts for fractionation between both (a) CaCO 3 and the CO 3 2− pool (α c/CO 3 2- ), and (b) CO 3 2− and water (α CO 3 2- /w ), as a function of temperature, pH, salinity, calcite saturation state (Ω), the residence time of the dissolved inorganic carbon (DIC) in solution, and the activity of the enzyme carbonic anhydrase. The model results suggest that: (1) The equilibrium α c/w is only approached in solutions with low Ω (i.e. close to 1) and low ionic strength such as in the cave system of Devils Hole, Nevada. (2) The sensitivity of α c/w to the solution pH and/or the mineral growth rate depends on the level of isotopic equilibration between the CO 3 2− pool and water. When the CO 3 2− pool approaches isotopic equilibrium with water, small negative pH and/or growth rate effects on α c/w of about 1–2‰ occur where these parameters covary with Ω. In contrast, isotopic disequilibrium between CO 3 2− and water leads to strong ( > 2‰) positive or negative pH and growth rate effects on α CO 3 2- /w (and α c/w ) due to the isotopic imprint of oxygen atoms derived from HCO 3 − , CO 2 , H 2 O and/or OH − . (3) The temperature sensitivity of α c/w originates from the negative effect of temperature on α CO 3 2- /w and is expected to deviate from the commonly accepted value (−0.22 ± 0.02‰/°C between 0 and 30 °C; Kim and O'Neil, 1997) when the CO 3 2− pool is not in isotopic equilibrium with water. (4) The model suggests that the δ 18 O of planktic and benthic foraminifers reflects a quantitative precipitation of DIC in isotopic equilibrium with a high-pH calcifying fluid, leading to a relatively constant foraminifer calcite δ 18 O-temperature relationship (−0.21 ± 0.01‰/°C). The lower average coral δ 18 O data relative to foraminifers and other calcifiers is best explained by the precipitation of internal DIC derived from hydrated CO 2 in a high-pH calcifying fluid and minimal subsequent DIC-H 2 O isotopic equilibration. This leads to a reduced and variable coral aragonite δ 18 O-temperature relationship (−0.11 to −0.22‰/°C). Together, the model presented here reconciles observations of oxygen isotope fractionation over a range of CaCO 3 -DIC-H 2 O systems.