The Marine20 marine radiocarbon age calibration curve provides a non-polar global-average record of radiocarbon from 0–55 cal kBP, serving as a baseline for regional oceanic variation. It is intended for calibration of marine radiocarbon samples from non-polar regions but not for polar regions where variability in sea ice, ocean upwelling, and air-sea gas exchange may affect marine radiocarbon concentrations. Marine20 is based on 500 simulations using a box model of the global carbon cycle, forced by posterior realizations of the Northern Hemispheric IntCal20 curve and reconstructed CO₂ changes from ice cores. This allows for the incorporation of carbon cycle dynamics and temporal changes in atmospheric ¹⁴C levels. The box model simulations of the global-average marine radiocarbon reservoir age are similar to those of a more complex three-dimensional ocean general circulation model. However, the simplicity and speed of the box model allow for a Monte Carlo approach to rigorously propagate uncertainty through the final Marine20 curve. This robust propagation of uncertainty is fundamental to providing reliable precision for marine radiocarbon age calibration. The updated values for ΔR, the regional marine radiocarbon reservoir age corrections required to calibrate against Marine20, can be found at the database http://calib.org/marine/. The Marine20 curve is constructed using the BICYCLE global carbon cycle model, which incorporates a globally averaged atmospheric box and modules of the terrestrial and oceanic components of the carbon cycle. The model is driven by temporal changes in boundary conditions and simulates changes in the carbon cycle, including ¹³C and ¹⁴C. The use of BICYCLE and time-dependent forcings allows for more complex and realistic inclusion of key carbon cycle changes extending back to 55 cal kBP. Improvements in the Bayesian modeling of IntCal20 allow for the first time to rigorously incorporate the time-dependent uncertainty of atmospheric ¹⁴C into the marine age calibration curve through the Monte-Carlo approach. Together, these features enable more accurate estimation of the global marine calibration curve and the globally averaged MRA, especially beyond the Holocene. The results from BICYCLE are almost indistinguishable from the three-dimensional Large Scale Geostrophic Ocean General Circulation Model (LSG OGCM) and agree well with pre-bomb data based on seawater and marine shells. The paper discusses the requirements for a computer model suitable for radiocarbon age calibration, how uncertainty in driving inputs can be propagated through such a model, and provides a tutorial on quantifying output uncertainty using Monte-Carlo techniques. It also justifies the recommended approach to marine calibration using local ΔR adjustments based on observations rather than directly using regional output from more complex three-dimensional models. The paper presents a short summary of BICYCLE and describes the key processes/forcings and their uncertainties propagated through to the final calibration curveThe Marine20 marine radiocarbon age calibration curve provides a non-polar global-average record of radiocarbon from 0–55 cal kBP, serving as a baseline for regional oceanic variation. It is intended for calibration of marine radiocarbon samples from non-polar regions but not for polar regions where variability in sea ice, ocean upwelling, and air-sea gas exchange may affect marine radiocarbon concentrations. Marine20 is based on 500 simulations using a box model of the global carbon cycle, forced by posterior realizations of the Northern Hemispheric IntCal20 curve and reconstructed CO₂ changes from ice cores. This allows for the incorporation of carbon cycle dynamics and temporal changes in atmospheric ¹⁴C levels. The box model simulations of the global-average marine radiocarbon reservoir age are similar to those of a more complex three-dimensional ocean general circulation model. However, the simplicity and speed of the box model allow for a Monte Carlo approach to rigorously propagate uncertainty through the final Marine20 curve. This robust propagation of uncertainty is fundamental to providing reliable precision for marine radiocarbon age calibration. The updated values for ΔR, the regional marine radiocarbon reservoir age corrections required to calibrate against Marine20, can be found at the database http://calib.org/marine/. The Marine20 curve is constructed using the BICYCLE global carbon cycle model, which incorporates a globally averaged atmospheric box and modules of the terrestrial and oceanic components of the carbon cycle. The model is driven by temporal changes in boundary conditions and simulates changes in the carbon cycle, including ¹³C and ¹⁴C. The use of BICYCLE and time-dependent forcings allows for more complex and realistic inclusion of key carbon cycle changes extending back to 55 cal kBP. Improvements in the Bayesian modeling of IntCal20 allow for the first time to rigorously incorporate the time-dependent uncertainty of atmospheric ¹⁴C into the marine age calibration curve through the Monte-Carlo approach. Together, these features enable more accurate estimation of the global marine calibration curve and the globally averaged MRA, especially beyond the Holocene. The results from BICYCLE are almost indistinguishable from the three-dimensional Large Scale Geostrophic Ocean General Circulation Model (LSG OGCM) and agree well with pre-bomb data based on seawater and marine shells. The paper discusses the requirements for a computer model suitable for radiocarbon age calibration, how uncertainty in driving inputs can be propagated through such a model, and provides a tutorial on quantifying output uncertainty using Monte-Carlo techniques. It also justifies the recommended approach to marine calibration using local ΔR adjustments based on observations rather than directly using regional output from more complex three-dimensional models. The paper presents a short summary of BICYCLE and describes the key processes/forcings and their uncertainties propagated through to the final calibration curve