A Living Benchmark for Machine Learning on Tabular Data 💫

---

🚀 Leaderboard	📂 Example Scripts	📊 Dataset Curation	📄 Paper

---

TabArena is a living benchmarking system that makes benchmarking tabular machine learning models a reliable experience. TabArena implements best practices to ensure methods are represented at their peak potential, including cross-validated ensembles, strong hyperparameter search spaces contributed by the method authors, early stopping, model refitting, parallel bagging, memory usage estimation, and more.

TabArena currently consists of:

• 51 manually curated tabular datasets representing real-world tabular data tasks.
• 9 to 30 evaluated splits per dataset.
• 27+ tabular machine learning methods, including 10+ tabular foundation models.
• More than 50 million trained models across the benchmark, with all validation and test predictions cached to enable tuning and post-hoc ensembling analysis.
• A live TabArena leaderboard showcasing the results.

⚡ Quickstart

Tip

The fastest way to try TabArena end-to-end:

pip install uv
git clone https://github.com/autogluon/tabarena.git && cd tabarena
uv venv --seed --python 3.12 && source .venv/bin/activate
uv pip install --prerelease=allow -e "./packages/tabarena[benchmark]"
python examples/benchmarking/run_quickstart_tabarena.py

For other install paths (eval-only, editable AutoGluon, dependency), see Installation below.

🕹️ Use Cases

We share more details on various use cases of TabArena in our examples:

• 📊 Benchmarking Predictive Machine Learning Models: please refer to examples/benchmarking.
• 🚀 Using SOTA Tabular Models Benchmarked by TabArena: please refer to examples/running_tabarena_models.
• 🧪 Advanced and Specialized Usage: please refer to examples/advanced.
• 🗃️ Analysing Metadata and Meta-Learning: please refer to examples/meta.
• 📈 Generating Plots and Leaderboards: please refer to examples/plots.
• 🔁 Reproducibility: we share instructions for reproducibility in examples.

Datasets

Please refer to our dataset curation repository to learn more about or contributed data!

More Documentation

TabArena code is currently being polished. Detailed Documentation for TabArena will be available soon.

🪄 Installation

Important

Requires Python 3.11–3.13 and uv.

TabArena is a uv workspace; its installable packages live under packages/ (tabarena, bencheval, tabflow_slurm). Install the tabarena package directly from packages/tabarena with the extras you need. The --prerelease=allow flag is required so uv resolves the pre-release dependency.

First clone the repo and create a virtual environment (one time):

git clone https://github.com/autogluon/tabarena.git
cd tabarena
uv venv --seed --python 3.12
source .venv/bin/activate

Then pick the install path that matches what you want to do:

📊 Evaluation only — leaderboards & metrics, no model fitting

uv pip install --prerelease=allow -e "./packages/tabarena"

🚀 Benchmark — core set of models for benchmarking

Installs the core models used for standard benchmarking: tabpfn, tabicl, ebm, search_spaces, realmlp, tabdpt, tabm.

uv pip install --prerelease=allow -e "./packages/tabarena[benchmark]"

➕ Benchmark + Extended — core models plus the extended model set

The extended extra is experimental and may fail to resolve or install due to incompatible version requirements across model dependencies. Use it only if you specifically need every model in a single environment; otherwise prefer benchmark or benchmark plus one specific model.

Layers the extended model set (modernnca, xrfm, sap-rpt-oss, ...) on top of the core benchmark set.

uv pip install --prerelease=allow -e "./packages/tabarena[benchmark,extended]"

To install only one extended model on top of benchmark (recommended over extended when you only need a single extra model), pass its extra by name — for example, just xrfm:

uv pip install --prerelease=allow -e "./packages/tabarena[benchmark,xrfm]"

🛠️ Developer — editable AutoGluon + editable TabArena

Create a virtual environment in your workspace directory (it spans both repos cloned below, so .venv lives at the workspace root rather than inside either repo):

uv venv --seed --python 3.12 .venv
source .venv/bin/activate

Install editable AutoGluon and TabArena:

git clone https://github.com/autogluon/autogluon.git
./autogluon/full_install.sh

git clone https://github.com/autogluon/tabarena.git
uv pip install --prerelease=allow -e "./tabarena/packages/tabarena[benchmark]"

In PyCharm, mark packages/tabarena/src/ and each autogluon/src/ subdirectory as Sources Root so imports resolve.

📦 Use TabArena as a dependency

Add the following to your project's dependencies:

"tabarena @ git+https://github.com/autogluon/tabarena.git#subdirectory=packages/tabarena"

📦 TabArena Artifacts

TabArena caches predictions, results, and leaderboards as downloadable artifacts so you can reproduce or extend any analysis without re-running the benchmark.

Artifact tiers, sizes, and examples

Artifacts download to ~/.cache/tabarena/ by default. Override the location with the TABARENA_CACHE environment variable.

Raw data is ~100 GB per method type. Point TABARENA_CACHE at a large disk before downloading it.

Tier	Contents	Size / method	Example
Raw data	Per-child test predictions, full metadata, system info	~100 GB	inspect_raw_data.py
Processed data	Minimal data for HPO simulation, portfolios, leaderboards	~10 GB	inspect_processed_data.py
Results	Per-config / HPO DataFrames (test error, val error, train time, inference time)	<1 MB	run_generate_main_leaderboard.py
Leaderboards	Aggregated ELO, win-rate, average rank, improvability	<1 MB	—
Figures & Plots	Generated from results and leaderboards	—	—

📄 Citation

Tip

If you use TabArena in a scientific publication, please cite our paper.

TabArena: A Living Benchmark for Machine Learning on Tabular Data Nick Erickson, Lennart Purucker, Andrej Tschalzev, David Holzmüller, Prateek Mutalik Desai, David Salinas, Frank Hutter NeurIPS 2025, Datasets and Benchmarks Track

📄 arXiv · 🎤 NeurIPS poster & video

BibTeX

The entry uses year=2026 because NeurIPS'25 proceedings are published in 2026.

@article{erickson2026tabarena,
  title   = {TabArena: A Living Benchmark for Machine Learning on Tabular Data},
  author  = {Erickson, Nick and Purucker, Lennart and Tschalzev, Andrej and Holzm{\"u}ller, David and Desai, Prateek and Salinas, David and Hutter, Frank},
  journal = {Advances in Neural Information Processing Systems},
  volume  = {38},
  year    = {2026}
}

---

Relation to TabRepo

TabArena was built upon and now replaces TabRepo. To see details about TabRepo, the portfolio simulation repository, refer to tabrepo.md.