The article introduces PartitionFinder, a new open-source program designed to objectively select the best partitioning schemes and nucleotide substitution models for phylogenetic analyses. Partitioning involves estimating independent models of molecular evolution for different sets of sites in a sequence alignment, and choosing an appropriate partitioning scheme is crucial for accurate phylogenetic reconstruction. Traditional methods often rely on biological intuition or ad hoc approaches, which can lead to overparameterization and suboptimal results. PartitionFinder addresses these issues by implementing two new objective methods that significantly outperform previous approaches, including ad hoc selection and hierarchical clustering. These methods allow for the comparison of millions of partitioning schemes in realistic time frames, making it feasible to select the best partitioning scheme even for large multilocus DNA data sets. The program uses information-theoretic metrics such as the Bayesian information criterion (BIC), Akaike information criterion (AIC), and corrected AIC (AICc) to guide the selection process. The authors demonstrate that PartitionFinder can find optimal partitioning schemes in all ten data sets tested, outperforming both exhaustive and heuristic searches. They also compare PartitionFinder's performance with commonly used a priori partitioning schemes, showing that these schemes are often suboptimal. Overall, PartitionFinder is expected to improve the accuracy of phylogenetic analyses by facilitating the objective selection of partitioning schemes.The article introduces PartitionFinder, a new open-source program designed to objectively select the best partitioning schemes and nucleotide substitution models for phylogenetic analyses. Partitioning involves estimating independent models of molecular evolution for different sets of sites in a sequence alignment, and choosing an appropriate partitioning scheme is crucial for accurate phylogenetic reconstruction. Traditional methods often rely on biological intuition or ad hoc approaches, which can lead to overparameterization and suboptimal results. PartitionFinder addresses these issues by implementing two new objective methods that significantly outperform previous approaches, including ad hoc selection and hierarchical clustering. These methods allow for the comparison of millions of partitioning schemes in realistic time frames, making it feasible to select the best partitioning scheme even for large multilocus DNA data sets. The program uses information-theoretic metrics such as the Bayesian information criterion (BIC), Akaike information criterion (AIC), and corrected AIC (AICc) to guide the selection process. The authors demonstrate that PartitionFinder can find optimal partitioning schemes in all ten data sets tested, outperforming both exhaustive and heuristic searches. They also compare PartitionFinder's performance with commonly used a priori partitioning schemes, showing that these schemes are often suboptimal. Overall, PartitionFinder is expected to improve the accuracy of phylogenetic analyses by facilitating the objective selection of partitioning schemes.