Atom probe tomography (APT) is routinely used for analyzing property-enhancing particles in the nanometer-size range and below, and plays a prominent role in the analysis of solute clusters. However, the question of how well these small particles are measured has never been addressed because of a lack of a reliable benchmark. Here, to address this critical gap, we use an approach that allows direct comparison of APT and small-angle (X-Ray) scattering (SA(X)S) performed on the same material. We introduce the notion of an effective spatial resolution for the analysis of particles, which, importantly in this context, is very different than the technique's inherent spatial resolution. This effective resolution is highly specific to the system being considered, as well as the analysis conditions. There is no hard limit below which the technique will fail, but for particles with a radius of order of 2σ = 1 nm, i.e. 250 atoms, cannot be accurately measured, even though the particles are detected. This thorough metrological assessment of APT in the analysis of particles allows us to discuss the pulse spread function of the technique and the physics underpinning its limits. We conclude that great care should be taken when analysing solute clusters by APT, in particular when reporting particle size and composition.