Contamination Map Methodology

Tabular / Model

Real-TabPFN-2.5

Organisation
Prior Labs
Family
TabPFN · version 2.5 (real-data fine-tuned)
Released
2025-11-11 (arXiv v1 submission date of the TabPFN-2.5 technical report https://arxiv.org/abs/2511.08667)
Checkpoints
Prior-Labs/tabpfn_2_5: tabpfn-v2.5-classifier-v2.5_default.ckpt, tabpfn-v2.5-classifier-v2.5_real.ckpt, tabpfn-v2.5-classifier-v2.5_real-large-features.ckpt, tabpfn-v2.5-classifier-v2.5_real-large-samples-and-features.ckpt, tabpfn-v2.5-regressor-v2.5_real.ckpt, tabpfn-v2.5-regressor-v2.5_real-variant.ckpt
Weights
Open · tabpfn-2.5-license-v1.1 (non-commercial)
Links
PaperModel cardCode

Training data disclosure: Full list

Level determined for 100% of benchmark cells.

“It is trained on a curated corpus of 43 real-world tabular datasets sourced from OpenML and Kaggle, deduplicated against all internal benchmarks and the full TabArena suite.”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“The following table lists the datasets curated for fine-tuning, along with their sources and access links.”
arxiv.org · paper · Appendix C.1 · retrieved 2026-10-15
“Real-TabPFN-2.5: continued pre-training on real-world datasets (for details please see Appendix C.1 of the model tech report).”
huggingface.co · model card · Training Data and Priors · retrieved 2026-10-15

Other declared training data

Declared exclusions

Benchmark cells

TabArena

Not seen airfoil_self_noise · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen Amazon_employee_access · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen anneal · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen Another-Dataset-on-used-Fiat-500 · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen APSFailure · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen bank-marketing · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen Bank_Customer_Churn · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen Bioresponse · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen blood-transfusion-service-center · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen churn · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen coil2000_insurance_policies · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen concrete_compressive_strength · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen credit-g · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen credit_card_clients_default · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen customer_satisfaction_in_airline · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen diabetes · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen Diabetes130US · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen diamonds · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen E-CommereShippingData · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen Fitness_Club · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen Food_Delivery_Time · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen GiveMeSomeCredit · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen hazelnut-spread-contaminant-detection · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen healthcare_insurance_expenses · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen heloc · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen hiva_agnostic · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen houses · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen HR_Analytics_Job_Change_of_Data_Scientists · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen in_vehicle_coupon_recommendation · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen Is-this-a-good-customer · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen kddcup09_appetency · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen Marketing_Campaign · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen maternal_health_risk · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen miami_housing · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen MIC · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen NATICUSdroid · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen online_shoppers_intention · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen physiochemical_protein · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen polish_companies_bankruptcy · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen qsar-biodeg · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen QSAR-TID-11 · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen QSAR_fish_toxicity · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen SDSS17 · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen seismic-bumps · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen splice · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen students_dropout_and_academic_success · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen superconductivity · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen taiwanese_bankruptcy_prediction · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen website_phishing · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen wine_quality · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Not seen jm1 · the authors state the dataset was excluded from training

Evidence (2)
“deduplicated against all internal benchmarks and the full TabArena suite”
arxiv.org · paper · Section 3, Data · retrieved 2026-10-15
“the process was extended to deduplicate the training datasets against all internal benchmarks, our curated in-house validation suite, and the public TabArena benchmark”
arxiv.org · paper · Appendix C · retrieved 2026-10-15

Notes

TabPFN-2.5 fine-tuned on 43 real datasets (15 OpenML + 28 Kaggle in Table C.1; the subset of Real-TabPFN's 71 that survived the TabArena deduplication). Marked with a globe emoji on the model card; this is the default classification checkpoint of the 2.5 release and the 'RealTabPFN-v2.5' entry on TabArena.

Sources