Consistent analysis of the Atlantic Meridional Overturning Circulation (AMOC) in climate model simulations and observations is important to advance our understanding of AMOC trends, variability, their latitudinal coherency, and related mechanisms. The concept of such a common framework into which both observations and models can be mapped and subsequently analyzed has emerged under the term AMOC Metrics. Such metrics further facilitate the joint interpretation of models and data, and promote objectivity in model intercomparisons. As a key task of our AMOC Metrics Project, we have created new tools to compute AMOC transports from models as consistently as possible to their observational counterparts. We focus on AMOC transport estimates from both the RAPID Array at 26°N and the Meridional Overturning Variability Experiment (MOVE) Array at 16°N. Using new tools, we intercompare trends and variability in both simulated and observed transports at these two latitudes. Our analysis reveals significant sensitivity of transports to reference level details and how the barotropic mode is calculated and included in the transport estimates, confirming prior findings from observational-based analysis. We also investigate how well the level of no-motion assumption holds at interannual-to-decadal timescales. These seemly minor details can significantly impact AMOC trends and variability, leading to a lack of latitudinal coherency between the two latitudes. Our main results are not qualitatively sensitive to the resolution of the ocean model, but there are some differences in detail, and care needs to be taken in defining geographic regimes, e.g., Florida Current vs. interior ocean, when the ocean current structures have very different widths at different resolutions.