LiTS-CryoBench is a benchmark dataset constructed for personalized liver tumor cryoablation planning. It is derived from the publicly available LiTS dataset, with a curated subset of cases selected to support planning-oriented research rather than pure segmentation tasks. The dataset consists of contrast-enhanced abdominal CT scans with corresponding liver and tumor annotations inherited from LiTS. Based on these cases, additional preprocessing and filtering steps were performed to ensure suitability for cryoablation planning, including quality control of tumor visibility, spatial coverage of the liver, and anatomical completeness. All data have been anonymized and formatted to facilitate downstream algorithm development. Unlike conventional liver tumor datasets that focus primarily on detection or segmentation, LiTS-CryoBench is specifically designed to serve as a benchmark for cryoablation planning. It supports research on probe placement optimization, ablation zone modeling, safety margin assessment, and planning evaluation under realistic anatomical constraints. The dataset can be used to develop and compare methods for image-guided cryoablation planning, simulation, and decision support. LiTS-CryoBench is intended for research and non-commercial use only and aims to provide a standardized evaluation platform for reproducible comparison of liver tumor cryoablation planning methods.