The Atlantic meridional overturning circulation (AMOC) plays an important role in regulating Earth’s climate. Concerns over its stability in a warming climate have prompted a concerted effort to monitor the AMOC in recent decades, most prominently with the RAPID array at 26N. Yet, such direct observational records are still too short to discriminate between natural decadal variability and long-term decline. This has motivated the development and use of indirect AMOC proxies, based more readily observed quantities such as sea surface temperature and coastal sea level for which long-term records do exist. However, these proxies are founded on correlations and plausible physical relationship rather than robustly established dynamical links. Moreover, relationships may not be exclusive, with factors other than AMOC variability also driving changes in the proxy variables (e.g., sea level or SST), with relative contributions of AMOC and non-AMOC factors potentially also being timescale dependent. The CMIP6 data provide an ideal opportunity to test the robustness of AMOC proxies, the aim of this project. First, the zonally-integrated meridional transport (ZIMT) will be calculated from the model velocity fields and various statistical approaches used to characterise ZIMT variability. Next, a range of commonly used AMOC proxies will be diagnosed from the model output and the strength of their relationship with the ZIMT assessed. Novel proxies may also be sought and tested. We may also consider whether there is a relationship between a model’s ability to represent a proxy quantity and the realism of its AMOC judged, as far as is possible, against the RAPID array data. Finally, based on the strength of the proxies we will attempt to assess the most likely behaviour of the AMOC over the last century.