Prior Labs Releases TabPFN-3.5: A Tabular Foundation Model That Beats the Winning Otto Kaggle Solution With Default Settings

Prior Labs launched TabPFN-3.5, a tabular foundation model that outperforms the winning 2015 Otto Kaggle solution and tops seven tabular benchmarks without dataset-specific tuning. It achieves superior accuracy using only synthetic pretraining and default settings, running efficiently on a single GPU.