---
modelA: "TabPFN v3"
modelB: "Google TabFM"
canonicalSlug: "tabpfn-vs-tabfm"
winsA: 5
winsB: 14
ties: 2
winRateA: "23.8%"
winRateB: "66.7%"
eloA: "936 (±59)"
eloB: "1111 (±101)"
avgLatencyA: "12.52"
avgLatencyB: "742.64"
canonicalUrl: "https://carlahq.eu/datasets/comparison/tabpfn-vs-tabfm"
---

# TabPFN vs TabFM Tabular Benchmark Comparison

> Empirical head-to-head evaluation across Carla's open-source tabular foundation model benchmark suite.

## Head-to-Head Scoreboard
| Metric | TabPFN v3 | Google TabFM | Comparison Result |
| :--- | :--- | :--- | :--- |
| **Head-to-Head Wins** | **5** (23.8%) | 14 (66.7%) | TabFM leads (2 ties) |
| **Bradley-Terry Elo Rating** | **936 (±59)** | 1111 (±101) | Global tournament rating across repeated splits |
| **Improvability Regret** | 1.69% | 0.85% | Normalized average error against dataset optimal ceilings |
| **Average Inference Latency** | 12.52 | 742.64 | Measured per-split forward-pass timing |
| **Browser Execution** | Server-side only (Requires Python runtime / CUDA / PyTorch) | Server-side only (Requires Python / JAX / PyTorch environment) | — |
| **Open Source License** | Research & Non-Commercial | Open Source (Apache 2.0) | — |

## Architectural Specifications
| Dimension | TabPFN v3 | Google TabFM |
| :--- | :--- | :--- |
| **Organization** | Prior Labs | Google Research |
| **Architecture** | Prior-Data Fitted Network (PFN) with Causally Invariant Embeddings | Columnar Embedding & Self-Supervised In-Context Transformer |
| **Pretraining Prior** | Bayesian structural equation models and synthetic Gaussian priors | Generative synthetic distributions and masked column modeling |
| **Context Window** | Fixed context window (typically 1k–10k training samples) | Columnar token budget per batch |
| **Execution Engine** | Server-side only (Requires Python runtime / CUDA / PyTorch) | Server-side only (Requires Python / JAX / PyTorch environment) |
| **Repository** | [TabPFN Code](https://github.com/PriorLabs/tabpfn) | [TabFM Code](https://github.com/google-research/tabfm) |
| **Paper / ArXiv** | [TabPFN Paper](https://arxiv.org/abs/2207.01848) | [TabFM Paper](https://github.com/google-research/tabfm) |

## Dataset-by-Dataset Benchmark Results
| Dataset | Business Domain | Task Type | Metric | TabPFN Score | TabFM Score | Delta (%) | Winner |
| :--- | :--- | :--- | :--- | :--- | :--- | :--- | :--- |
| [Abalone](/datasets/undefined.md) | Agriculture, Forestry & Fishing | `regression` | undefined | 2.03 | 2.02 | +0.0% | **TabFM** |
| [Adult](/datasets/undefined.md) | Economics & Public Policy | `binary_classification` | undefined | 0.0799 | 0.0677 | +0.0% | **TabFM** |
| [Airfoil Self Noise](/datasets/undefined.md) | UCI | `regression` | undefined | 1.02 | 1.11 | -0.1% | **TabPFN** |
| [Amazon Employee Access](/datasets/undefined.md) | Information Technology & Enterprise Security | `binary_classification` | undefined | 0.1542 | 0.1273 | +0.0% | **TabFM** |
| [Bank Marketing](/datasets/undefined.md) | Finance & Banking | `binary_classification` | undefined | 0.1883 | 0.1840 | +0.0% | **TabFM** |
| [Blood Transfusion Service Center](/datasets/undefined.md) | Healthcare & Biomedicine | `binary_classification` | undefined | 0.2462 | 0.2445 | +0.0% | **TabFM** |
| [Breast W](/datasets/undefined.md) | Healthcare & Life Sciences | `binary_classification` | undefined | 0.0056 | 0.0055 | +0.0% | **Tie** |
| [Car](/datasets/undefined.md) | Automotive & Fleet Management | `multiclass_classification` | undefined | 0.0238 | 0.0108 | +0.0% | **TabFM** |
| [Compas Two Years](/datasets/undefined.md) | Legal & Public Safety | `binary_classification` | undefined | 0.2615 | 0.2619 | -0.0% | **TabPFN** |
| [Credit G](/datasets/undefined.md) | Finance & Banking | `binary_classification` | undefined | 0.2031 | 0.1938 | +0.0% | **TabFM** |
| [Diabetes](/datasets/undefined.md) | Healthcare & Biomedicine | `binary_classification` | undefined | 0.1607 | 0.1623 | -0.0% | **TabPFN** |
| [Employee Salaries](/datasets/undefined.md) | Human Resources & Workforce Analytics | `regression` | undefined | 8,308 | 10,435 | -2126.6% | **TabPFN** |
| [Fitness Club](/datasets/undefined.md) | Fitness, Sports & Recreation | `binary_classification` | undefined | 0.1792 | 0.1795 | -0.0% | **TabPFN** |
| [House Sales](/datasets/undefined.md) | Real Estate & Property Valuation | `regression` | undefined | 102,039 | 98,887 | +3152.6% | **TabFM** |
| [Houses](/datasets/undefined.md) | Real Estate & Urban Planning | `regression` | undefined | 39,523 | 37,481 | +2041.8% | **TabFM** |
| [Monks Problems 2](/datasets/undefined.md) | Data Science & Artificial Intelligence | `binary_classification` | undefined | 0.0000 | 0.0000 | 0.0% | **Tie** |
| [Phoneme](/datasets/undefined.md) | Speech Processing & Acoustic Engineering | `binary_classification` | undefined | 0.0274 | 0.0202 | +0.0% | **TabFM** |
| [Spambase](/datasets/undefined.md) | Cybersecurity & IT Infrastructure | `binary_classification` | undefined | 0.0085 | 0.0066 | +0.0% | **TabFM** |
| [Telco Customer Churn](/datasets/undefined.md) | Telecommunications & Subscription Services | `binary_classification` | undefined | 0.1486 | 0.1474 | +0.0% | **TabFM** |
| [Titanic](/datasets/undefined.md) | Maritime Safety & Actuarial Risk Analysis | `binary_classification` | undefined | 0.1360 | 0.1166 | +0.0% | **TabFM** |
| [Vehicle](/datasets/undefined.md) | Automotive Engineering & Computer Vision | `multiclass_classification` | undefined | 0.2263 | 0.1986 | +0.0% | **TabFM** |

## Canonical Resources
- **HTML Benchmark Page:** [https://carlahq.eu/datasets/comparison/tabpfn-vs-tabfm](https://carlahq.eu/datasets/comparison/tabpfn-vs-tabfm)
- **Carla Tabular Suite:** [https://carlahq.eu/datasets/](https://carlahq.eu/datasets/)