---
modelA: "TabICLv2"
modelB: "Kumo Tabular Small"
canonicalSlug: "tabiclv2-vs-kumo"
winsA: 6
winsB: 14
ties: 1
winRateA: "28.6%"
winRateB: "66.7%"
eloA: "916 (±60)"
eloB: "1042 (±50)"
avgLatencyA: "4.26"
avgLatencyB: "14.52"
canonicalUrl: "https://carlahq.eu/datasets/comparison/tabiclv2-vs-kumo"
---

# TabICLv2 vs Kumo Tabular Small Tabular Benchmark Comparison

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

## Head-to-Head Scoreboard
| Metric | TabICLv2 | Kumo Tabular Small | Comparison Result |
| :--- | :--- | :--- | :--- |
| **Head-to-Head Wins** | **6** (28.6%) | 14 (66.7%) | Kumo Tabular Small leads (1 ties) |
| **Bradley-Terry Elo Rating** | **916 (±60)** | 1042 (±50) | Global tournament rating across repeated splits |
| **Improvability Regret** | 2.13% | 1.23% | Normalized average error against dataset optimal ceilings |
| **Average Inference Latency** | 4.26 | 14.52 | Measured per-split forward-pass timing |
| **Browser Execution** | Native WebGPU Browser Runtime (Carla Engine / ONNX Runtime WASM) | Native WebGPU & WASM Browser Runtime (via ONNX Runtime) | — |
| **Open Source License** | Open Source (BSD-3-Clause) | Open Source (Apache 2.0) | — |

## Architectural Specifications
| Dimension | TabICLv2 | Kumo Tabular Small |
| :--- | :--- | :--- |
| **Organization** | Inria SODA | Kumo AI / NVIDIA |
| **Architecture** | In-Context Tabular Transformer with Prefix Attention & KV Caching | In-Context Tabular Foundation Model with Cross-Attention |
| **Pretraining Prior** | Synthetic Copula-based Quantile Priors & Prior.py DAGs | Relational & Structured Data Priors (SDM Recipe) |
| **Context Window** | 1,000+ context rows with zero fine-tuning required | Configurable context rows with tiled row & ICL inference |
| **Execution Engine** | Native WebGPU Browser Runtime (Carla Engine / ONNX Runtime WASM) | Native WebGPU & WASM Browser Runtime (via ONNX Runtime) |
| **Repository** | [TabICLv2 Code](https://github.com/soda-inria/tabicl) | [Kumo Tabular Small Code](https://huggingface.co/nvidia/Kumo-Tabular) |
| **Paper / ArXiv** | [TabICLv2 Paper](https://arxiv.org/abs/2602.11139) | [Kumo Tabular Small Paper](https://arxiv.org/abs/2502.05324) |

## Dataset-by-Dataset Benchmark Results
| Dataset | Business Domain | Task Type | Metric | TabICLv2 Score | Kumo Tabular Small Score | Delta (%) | Winner |
| :--- | :--- | :--- | :--- | :--- | :--- | :--- | :--- |
| [Abalone](/datasets/undefined.md) | Agriculture, Forestry & Fishing | `regression` | undefined | 2.03 | 2.02 | +0.0% | **Kumo Tabular Small** |
| [Adult](/datasets/undefined.md) | Economics & Public Policy | `binary_classification` | undefined | 0.0793 | 0.0677 | +0.0% | **Kumo Tabular Small** |
| [Airfoil Self Noise](/datasets/undefined.md) | UCI | `regression` | undefined | 1.12 | 1.02 | +0.1% | **Kumo Tabular Small** |
| [Amazon Employee Access](/datasets/undefined.md) | Information Technology & Enterprise Security | `binary_classification` | undefined | 0.1483 | 0.1392 | +0.0% | **Kumo Tabular Small** |
| [Bank Marketing](/datasets/undefined.md) | Finance & Banking | `binary_classification` | undefined | 0.1970 | 0.1880 | +0.0% | **Kumo Tabular Small** |
| [Blood Transfusion Service Center](/datasets/undefined.md) | Healthcare & Biomedicine | `binary_classification` | undefined | 0.2446 | 0.2452 | -0.0% | **TabICLv2** |
| [Breast W](/datasets/undefined.md) | Healthcare & Life Sciences | `binary_classification` | undefined | 0.0052 | 0.0055 | -0.0% | **TabICLv2** |
| [Car](/datasets/undefined.md) | Automotive & Fleet Management | `multiclass_classification` | undefined | 0.0231 | 0.0253 | -0.0% | **TabICLv2** |
| [Compas Two Years](/datasets/undefined.md) | Legal & Public Safety | `binary_classification` | undefined | 0.2701 | 0.2617 | +0.0% | **Kumo Tabular Small** |
| [Credit G](/datasets/undefined.md) | Finance & Banking | `binary_classification` | undefined | 0.2016 | 0.1956 | +0.0% | **Kumo Tabular Small** |
| [Diabetes](/datasets/undefined.md) | Healthcare & Biomedicine | `binary_classification` | undefined | 0.1632 | 0.1621 | +0.0% | **Kumo Tabular Small** |
| [Employee Salaries](/datasets/undefined.md) | Human Resources & Workforce Analytics | `regression` | undefined | 8,126 | 9,029 | -903.3% | **TabICLv2** |
| [Fitness Club](/datasets/undefined.md) | Fitness, Sports & Recreation | `binary_classification` | undefined | 0.1795 | 0.1792 | +0.0% | **Kumo Tabular Small** |
| [House Sales](/datasets/undefined.md) | Real Estate & Property Valuation | `regression` | undefined | 111,524 | 98,512 | +13011.2% | **Kumo Tabular Small** |
| [Houses](/datasets/undefined.md) | Real Estate & Urban Planning | `regression` | undefined | 40,971 | 37,670 | +3300.8% | **Kumo Tabular Small** |
| [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 | 0.0246 | +0.0% | **Kumo Tabular Small** |
| [Spambase](/datasets/undefined.md) | Cybersecurity & IT Infrastructure | `binary_classification` | undefined | 0.0075 | 0.0073 | +0.0% | **Kumo Tabular Small** |
| [Telco Customer Churn](/datasets/undefined.md) | Telecommunications & Subscription Services | `binary_classification` | undefined | 0.1488 | 0.1595 | -0.0% | **TabICLv2** |
| [Titanic](/datasets/undefined.md) | Maritime Safety & Actuarial Risk Analysis | `binary_classification` | undefined | 0.1229 | 0.1276 | -0.0% | **TabICLv2** |
| [Vehicle](/datasets/undefined.md) | Automotive Engineering & Computer Vision | `multiclass_classification` | undefined | 0.2401 | 0.2244 | +0.0% | **Kumo Tabular Small** |

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