TabPFN

What it does

TabPFN is a pre-trained model for prediction tasks on structured data. Users provide a tabular dataset and a target; it can return class predictions and probabilities, continuous estimates with uncertainty, forecasts, anomaly scores, or synthetic tabular data. Documentation describes in-context inference without retraining for each dataset. A documented Python example uses the TabPFN classifier with a scikit-learn workflow and calls predict and predict_proba.

The product is aimed at teams evaluating or building prediction workflows, as well as academic researchers. Prior Labs lists use cases across finance, healthcare, industrial operations, energy, and technology. TabPFN can be tried in a web playground, used through a Python SDK or REST API, or deployed on private cloud or on-premises infrastructure; MCP is also listed. The free API has daily and monthly limits, and requests are rejected until a limit resets. Published model weights are offered under a non-commercial license; commercial private-cloud or on-premises deployment requires a license.

What it can do

• Binary and multiclass classification with calibrated probabilities
• Regression with uncertainty-aware outputs
• Time-series forecasting
• Anomaly detection
• Synthetic tabular data generation
• Fine-tuning

Works with

• scikit-learn interface
• Python SDK
• REST API
• Model Context Protocol (MCP)
• Amazon SageMaker
• Azure ML

Before you choose

• Free API access has daily and monthly limits; requests are rejected until the relevant limit resets.
• Published model weights carry non-commercial restrictions; commercial private-cloud or on-premises deployment requires a license. Free API use has separate terms.
• Paid plans are scoped to workload and requirements, but numeric rates and exact free API limits are not stated in the pricing page.

Listed under

• Data
• Analytics and BI
• Churn prediction
• Data science
• Enterprise
• Forecasting
• Fraud detection
• Predictive analytics

Get started

The free API plan lists an API-key signup option; a playground is also offered.

1. Choose the API, private-cloud, or on-premises option appropriate to your use.
2. For API access, get an API key; for open-source on-premises evaluation, start with the tabpfn package.
3. Prepare a structured dataset and target, then follow the Python quickstart or use the REST API.

Free API use is limited daily and monthly. Pro and Max API paid plans have no numeric rates stated; contact sales. Commercial private-cloud and on-premises use requires a license. Published model weights are for non-commercial use.

Source: priorlabs.ai/pricing, checked 1 Nov 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 1 Nov 2026. Each source supports part of the overview; availability and prices can change.

• TabPFN 3.5 | Prior Labs ↗priorlabs.ai
• Overview - Prior Labs ↗docs.priorlabs.ai
• Pricing | Prior Labs ↗priorlabs.ai
• Prior Labs ↗priorlabs.ai
• About | Prior Labs ↗priorlabs.ai
• Contact ↗priorlabs.ai

Company website

priorlabs.ai: product, pricing and policy pages

YC directory mirror

Batch and company name

Web search

Confirms the official site

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

manual_review

Last checked 1 Nov 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.