---
modelA: "TabICLv2"
modelB: "TabPFN v3"
canonicalSlug: "tabiclv2-vs-tabpfn"
winsA: 9
winsB: 11
ties: 1
winRateA: "42.9%"
winRateB: "52.4%"
eloA: "932 (±61)"
eloB: "936 (±60)"
avgLatencyA: "4.26"
avgLatencyB: "12.52"
canonicalUrl: "https://carlahq.eu/datasets/comparison/tabiclv2-vs-tabpfn"
---

# TabICLv2 vs TabPFN Tabular Benchmark Comparison

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

## Head-to-Head Scoreboard
| Metric | TabICLv2 | TabPFN v3 | Comparison Result |
| :--- | :--- | :--- | :--- |
| **Head-to-Head Wins** | **9** (42.9%) | 11 (52.4%) | TabPFN leads (1 ties) |
| **Bradley-Terry Elo Rating** | **932 (±61)** | 936 (±60) | Global tournament rating across repeated splits |
| **Improvability Regret** | 2.11% | 1.69% | Normalized average error against dataset optimal ceilings |
| **Average Inference Latency** | 4.26 | 12.52 | Measured per-split forward-pass timing |
| **Browser Execution** | Native WebGPU Browser Runtime (Carla Engine / ONNX Runtime WASM) | Server-side only (Requires Python runtime / CUDA / PyTorch) | — |
| **Open Source License** | Open Source (BSD-3-Clause) | Research & Non-Commercial | — |

## Architectural Specifications
| Dimension | TabICLv2 | TabPFN v3 |
| :--- | :--- | :--- |
| **Organization** | Inria SODA | Prior Labs |
| **Architecture** | In-Context Tabular Transformer with Prefix Attention & KV Caching | Prior-Data Fitted Network (PFN) with Causally Invariant Embeddings |
| **Pretraining Prior** | Synthetic Copula-based Quantile Priors & Prior.py DAGs | Bayesian structural equation models and synthetic Gaussian priors |
| **Context Window** | 1,000+ context rows with zero fine-tuning required | Fixed context window (typically 1k–10k training samples) |
| **Execution Engine** | Native WebGPU Browser Runtime (Carla Engine / ONNX Runtime WASM) | Server-side only (Requires Python runtime / CUDA / PyTorch) |
| **Repository** | [TabICLv2 Code](https://github.com/soda-inria/tabicl) | [TabPFN Code](https://github.com/PriorLabs/tabpfn) |
| **Paper / ArXiv** | [TabICLv2 Paper](https://arxiv.org/abs/2602.11139) | [TabPFN Paper](https://arxiv.org/abs/2207.01848) |

## Dataset-by-Dataset Benchmark Results
| Dataset | Business Domain | Task Type | Metric | TabICLv2 Score | TabPFN Score | Delta (%) | Winner |
| :--- | :--- | :--- | :--- | :--- | :--- | :--- | :--- |
| [Abalone](/datasets/undefined.md) | Agriculture, Forestry & Fishing | `regression` | undefined | 2.03 | 2.03 | +0.0% | **TabPFN** |
| [Adult](/datasets/undefined.md) | Economics & Public Policy | `binary_classification` | undefined | 0.0793 | 0.0799 | -0.0% | **TabICLv2** |
| [Airfoil Self Noise](/datasets/undefined.md) | UCI | `regression` | undefined | 1.12 | 1.02 | +0.1% | **TabPFN** |
| [Amazon Employee Access](/datasets/undefined.md) | Information Technology & Enterprise Security | `binary_classification` | undefined | 0.1483 | 0.1542 | -0.0% | **TabICLv2** |
| [Bank Marketing](/datasets/undefined.md) | Finance & Banking | `binary_classification` | undefined | 0.1970 | 0.1883 | +0.0% | **TabPFN** |
| [Blood Transfusion Service Center](/datasets/undefined.md) | Healthcare & Biomedicine | `binary_classification` | undefined | 0.2446 | 0.2462 | -0.0% | **TabICLv2** |
| [Breast W](/datasets/undefined.md) | Healthcare & Life Sciences | `binary_classification` | undefined | 0.0052 | 0.0056 | -0.0% | **TabICLv2** |
| [Car](/datasets/undefined.md) | Automotive & Fleet Management | `multiclass_classification` | undefined | 0.0231 | 0.0238 | -0.0% | **TabICLv2** |
| [Compas Two Years](/datasets/undefined.md) | Legal & Public Safety | `binary_classification` | undefined | 0.2701 | 0.2615 | +0.0% | **TabPFN** |
| [Credit G](/datasets/undefined.md) | Finance & Banking | `binary_classification` | undefined | 0.2016 | 0.2031 | -0.0% | **TabICLv2** |
| [Diabetes](/datasets/undefined.md) | Healthcare & Biomedicine | `binary_classification` | undefined | 0.1632 | 0.1607 | +0.0% | **TabPFN** |
| [Employee Salaries](/datasets/undefined.md) | Human Resources & Workforce Analytics | `regression` | undefined | 8,126 | 8,308 | -182.2% | **TabICLv2** |
| [Fitness Club](/datasets/undefined.md) | Fitness, Sports & Recreation | `binary_classification` | undefined | 0.1795 | 0.1792 | +0.0% | **TabPFN** |
| [House Sales](/datasets/undefined.md) | Real Estate & Property Valuation | `regression` | undefined | 111,524 | 102,039 | +9484.4% | **TabPFN** |
| [Houses](/datasets/undefined.md) | Real Estate & Urban Planning | `regression` | undefined | 40,971 | 39,523 | +1448.2% | **TabPFN** |
| [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 | 0.0278 | 0.0274 | +0.0% | **TabPFN** |
| [Spambase](/datasets/undefined.md) | Cybersecurity & IT Infrastructure | `binary_classification` | undefined | 0.0075 | 0.0085 | -0.0% | **TabICLv2** |
| [Telco Customer Churn](/datasets/undefined.md) | Telecommunications & Subscription Services | `binary_classification` | undefined | 0.1488 | 0.1486 | +0.0% | **TabPFN** |
| [Titanic](/datasets/undefined.md) | Maritime Safety & Actuarial Risk Analysis | `binary_classification` | undefined | 0.1229 | 0.1360 | -0.0% | **TabICLv2** |
| [Vehicle](/datasets/undefined.md) | Automotive Engineering & Computer Vision | `multiclass_classification` | undefined | 0.2401 | 0.2263 | +0.0% | **TabPFN** |

## Canonical Resources
- **HTML Benchmark Page:** [https://carlahq.eu/datasets/comparison/tabiclv2-vs-tabpfn](https://carlahq.eu/datasets/comparison/tabiclv2-vs-tabpfn)
- **Carla Tabular Suite:** [https://carlahq.eu/datasets/](https://carlahq.eu/datasets/)