Aim <p>Before conducting a clinical study involving pharmacometrics analyses using Non-Linear Mixed Effects Models (NLMEM), study design can be evaluated by computing the Fisher Information Matrix (FIM) via first-order model linearisation. A 2015 study showed that several dedicated tools provided consistent results, with block-diagonal FIM approximations aligning more closely with clinical trial simulations (CTS) than full matrix approaches. Since, some tools have evolved with new features like covariate modelling and inter-occasion variability (IOV), while others have been newly developed within software initially created for parameter estimation in NLMEM. Our first aim is to compare predictions from both older (PFIM, PopED) and newer tools (NONMEM$Design, MlxDesignEval and Pumas OptimalDesign), and secondly to compare available estimation software to derived empirical predictions from CTS.</p> Methods <p>The 2015 examples, involving an analytical pharmacokinetic (PK) model and a pharmacokinetic/pharmacodynamic model described by ODE, and new cases including a PK model with covariates and a cross over design with IOV were implemented in the five software. Predicted relative standard error (RSE) and D-criterion were computed. Empirical RSE and D-criterion were estimated from CTS with the different estimation tools.</p> Results <p>For a given FIM approximation, both RSE and D-criterion were consistent across evaluation software. Results derived from block diagonal FIM were more aligned with CTS, for which results were comparable across estimation tools. Predictions were especially reliable for covariate effects and appropriately reflected the order of magnitude of IOV variances.</p> Conclusion <p>Users can benefit from both reliable uncertainty prediction and parameter estimation regardless of their preferred software.</p>