Germline copy number variants (CNVs) play a significant role in hereditary diseases. However, the accurate detection of CNVs from targeted next-generation sequencing (NGS) gene panel data remains a challenging task. Several tools for calling CNVs within this context have been published to date, but the available benchmarks suffer from limitations, including testing on simulated data, testing on small datasets, and testing a small subset of published tools. In this work, we conducted a comprehensive benchmarking of 12 tools (Atlas-CNV, ClearCNV, ClinCNV, CNVkit, Cobalt, CODEX2, CoNVaDING, DECoN, ExomeDepth, GATK-gCNV, panelcn.MOPS, VisCap) on four validated gene panel datasets using their default parameters. We also assessed the impact of modifying 107 tool parameters and identified 13 parameter values that we suggest using to improve the tool F1 score. A total of 66 tool pair combinations were also evaluated to produce better meta-callers. Furthermore, we developed CNVbenchmarker2, a framework to help users perform their own evaluations. Our results indicated that in terms of F1 score, ClinCNV and GATK-gCNV were the best CNV callers. Regarding sensitivity, GATK-gCNV also exhibited particularly high performance. The results presented here provide an evaluation of the current state of the art in germline CNV detection from gene panel data and can be used as a reference resource when using any of the tools.