APBench (Astrodynamics Problem Bench) is a specialized benchmarking dataset for evaluating the performance, not for training, of Large Language Models in fundamental space engineering and astrodynamics problems. This dataset consists of numerical and analytical questions, designed to test models' comprehension and practical use for space engineering research and development.
