This article proposes a modification to the standard forcing–feedback diagnostic energy balance model to account for 1) differences between effective and equilibrium climate sensitivities and 2) the variation of effective sensitivity over time in climate change experiments with coupled atmosphere–ocean climate models. In the spirit of Hansen et al. an efficacy factor is applied to the ocean heat uptake. Comparing the time evolution of the surface warming in high and low efficacy models demonstrates the role of this efficacy in the transient response to CO₂ forcing. Abrupt CO₂ increase experiments show that the large efficacy of the Geophysical Fluid Dynamics Laboratory’s Climate Model version 2.1 (CM2.1) sets up in the first two decades following the increase in forcing. The use of an efficacy is necessary to fit this model’s global mean temperature evolution in periods with both increasing and stable forcing. The intermodel correlation of transient climate response with ocean heat uptake efficacy is greater than its correlation with equilibrium climate sensitivity in an ensemble of climate models used for the third and fourth Intergovernmental Panel on Climate Change (IPCC) assessments. When computed at the time of doubling in the standard experiment with 1% yr−1increase in CO₂, the efficacy is variable amongst the models but is generally greater than 1, averages between 1.3 and 1.4, and is as large as 1.75 in several models.