Google Research Introduces TabFM: A Zero-Shot Foundation Model for Tabular Data

Google Research has officially announced TabFM (Tabular Foundation Model), a zero-shot foundation model designed specifically for tabular data. The model can be applied directly to a wide range of tabular data tasks without requiring dataset-specific fine-tuning.

TabFM's core innovation lies in its zero-shot capability. Traditional tabular data models typically require large amounts of labeled data for training, but TabFM can generalize to unseen tabular datasets directly. This has broad application potential in data science and machine learning, especially in scenarios where labeled data is scarce.

The announcement was published on the Google Research Blog on June 30, 2026. Tabular data is one of the most common data types across enterprises and institutions, spanning finance, healthcare, e-commerce, and many other sectors.

This direction also reflects the trend of foundation models expanding from text and images to structured data. With TabFM, Google further solidifies its leadership in foundation model research.