Tripo AI 3D Generator
Mini toolsTabPFN
Overview
TabPFN is a tabular foundation model built by Prior Labs for classification and regression on structured datasets. It uses in-context learning to predict on new data without per-dataset training. The current default, TabPFN-3, supports datasets up to 1,000,000 rows and 200 features. It runs locally as an open source Python package with PyTorch and CUDA support, or via a hosted client.
About Prior Labs
Prior Labs is a Freiburg, Germany-based AI lab and pioneer of Tabular Foundation Models — pre-trained models for state-of-the-art predictions on structured data, no tuning needed. Its model TabPFN, published in Nature, is used in finance, healthcare, industrials, and energy. In May 2026, SAP agreed to acquire Prior Labs, planning to invest over €1B into it over four years while it operates independently.
