Abstract page for arXiv paper 2611.13001: One Table, Many Queries: Benchmarking Generalist Tabular Models Across Generation, Imputation, and Prediction