Climate change adaptation under resource constraints and future climate uncertainties would benefit from fully probabilistic climate risks assessments. Conducting such risk analyses requires assigning probabilities to the future greenhouse gases (GHG) and land-use scenarios used by global climate models. This paper proposes an approach to estimate the relative likelihood of carbon dioxide (CO2) concentration scenarios used in key climate change modeling experiments. The approach relies on the comparison of CO2 emissions from probabilistic simulations of Integrated Assessment Models (IAM) with compatible CO2 emissions diagnosed by global climate models participating in the Coupled Model Intercomparison Project Phase 5 (CMIP5) and 6 (CMIP6). The approach is demonstrated with five emission simulations from four IAMs, leading to independent estimates of the relative likelihood of CMIP5 Representation Concentration Pathways and CMIP6’ Shared Socioeconomic Pathways (SSP) up to 2100. Results suggest that SSP5-8.5 is an unlikely scenario for the second half of the century, but there is no clear consensus on the most likely scenario. Scenario likelihood is affected by a number of potential errors, including sampling errors, differences in emission sources simulated by the IAMs, and the lack of a common experimental framework for IAM simulations. These errors, along with the small IAM ensemble size, limit the applicability of the results. The delivery of fully probabilistic climate risk assessments would benefit from a coordinated probabilistic IAM experiment jointly designed with a coordinated climate modeling experiment where Earth System Model are driven by representative emission pathways.