# PriorLabs/TabPFN

⚡ TabPFN: Foundation Model for Tabular Data ⚡

Repository: https://github.com/PriorLabs/TabPFN
Canonical: https://ross.abutalabs.com/products/tabpfn
Homepage: http://priorlabs.ai
Language: Python
License: NOASSERTION
License Family: other
Topics: data-science, foundation-models, machine-learning, tabpfn, tabular-data
Last push: 2026-08-26T23:38:12+00:00

## Health v2 (maintenance only)
Score: 99/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 98, longevity 100
- inputs: {"age_days": 1524, "days_push": 7, "days_rel": 14, "gap_med": 6, "n_releases_24m": 26}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 7855, forks 781 (observed 2026-08-28T04:10:07.451920+00:00)

## What it is
TabPFN is a pre-trained tabular foundation model (transformer) that performs classification, regression, forecasting, and anomaly detection on structured data via in-context learning in a single forward pass, with a scikit-learn-compatible Python package. It is developed by Prior Labs, published in Nature, and also available through a hosted REST API and client SDK.

## Use cases
- predict on tabular datasets without training a model
- classify rows in a csv with calibrated probabilities
- regression on structured data with uncertainty estimates
- detect anomalies in tabular data
- forecast time series
- generate synthetic tabular data
- replace AutoML and tuned tree ensembles like XGBoost
- fit a model in seconds on small to medium datasets

## When to choose
- you need strong tabular predictions fast without hyperparameter tuning or pipeline building
- your datasets fit within model limits (up to ~1M rows with scaling mode, GPU recommended)
- you want a scikit-learn-compatible estimator that works out of the box
- you want state-of-the-art accuracy on structured data benchmarks

## When to avoid
- you need a permissively licensed library for commercial use without attribution provisions
- you have no GPU and very large datasets (CPU is limited to moderate sizes)
- you need full interpretability or custom model architectures
- you work primarily with unstructured data like text or images

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, data-science, llm-inference
- domain: machine-learning, data-science, artificial-intelligence
- platform: python, cross-platform
- tags: tabular-data, foundation-model, in-context-learning, classification, regression, anomaly-detection, time-series-forecasting, scikit-learn-compatible, transformer, prior-fitted-networks, gpu

## Member repositories
- PriorLabs/TabPFN (main) score 99

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:07.451920+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:33:47.782374+00:00, confidence not recorded.
  - readme: https://github.com/PriorLabs/TabPFN (fetched 2026-08-28T04:10:07.451920+00:00, sha 48f19818b65f)
  - homepage: http://priorlabs.ai (fetched 2026-08-29T08:30:27.071390+00:00, sha fe1cedd0914b)
  - site_page: https://priorlabs.ai/about (fetched 2026-08-29T08:30:27.085586+00:00, sha 0938a3e7c161)
  - site_page: https://docs.priorlabs.ai/overview (fetched 2026-08-29T08:30:27.087707+00:00, sha 8e9a3f7a46d7)
  - site_page: https://docs.priorlabs.ai/api-reference/getting-started (fetched 2026-08-29T08:30:27.089588+00:00, sha a7bb0855bc7a)
  - site_page: https://docs.priorlabs.ai/ (fetched 2026-08-29T08:30:27.082405+00:00, sha 5e4262efdb4b)
  - registry_pypi: https://pypi.org/pypi/tabpfn/json (fetched 2026-08-29T08:30:27.091397+00:00, sha 3eb3d8c90945)
  - site_page: https://priorlabs.ai/pricing (fetched 2026-08-29T08:30:27.080566+00:00, sha 7165d1b9f1c5)
  - site_page: https://priorlabs.ai/technical-reports (fetched 2026-08-29T08:30:27.084097+00:00, sha 9d28dd2cdd0e)
- Data as of 2026-08-30T08:39:29.467469+00:00.