Preview Program - TabPFN-3-Plus API
Who we are?
Structured data is the language of measurement and decision-making, yet AI still treats it as an afterthought.
At Prior Labs, we're building Multimodal Tabular Foundation Models (TFMs), starting with TabPFN, that understand tables natively, learning statistical reasoning directly from data.
What is this about?
TabPFN-3-Plus is a foundation model for tabular data - point it at your spreadsheet or database and it predicts outcomes (churn, fraud, prices, medical risk, demand, failures) without any training, tuning, or feature engineering. It's building on our TabPFN-2.5 model, which beats classic ML models like XGBoost and CatBoost on variety of tasks by far.
TabPFN-3-Plus deals with missing values, categorical values, skewed data, text features out-of-the-box - no data cleaning needed.
Get free API access a week before public release, with Prior Labs engineers on Discord to help you try it on datasets up to 1M rows, text columns included.
