Motivation: Model selection is a ubiquitous challenge in statistics. For penalized models, model selection typically entails tuning hyperparameters to maximize a measure of fit or minimize out-of-sample prediction error. However, these criteria fail to reflect other desirable characteristics, such as model sparsity, interpretability, or smoothness. Results: We present the R package pared for Pareto-based model selection using multi-objective optimization. The package provides a generic interface, pared_optimize(), that allows users to define model-specific tuning parameters, objective functions, and objective directions. Built-in wrappers are also provided for the elastic net, fused lasso, fused graphical lasso, and group graphical lasso. Gaussian process-based multi-objective optimization is used to approximate the Pareto front, and interactive graphics allow users to inspect trade-offs among fit, sparsity, smoothness, structural similarity, and other user-defined criteria. Availability and implementation: The pared R package and vignettes are available at https://github.com/priyamdas2/pared. The archived release corresponding to this manuscript is available on Zenodo with DOI: 10.5281/zenodo.20533535.