CatBoost
YandexGradient-boosted decision tree library from Yandex, known for handling categorical features natively and for ordered boosting, which reduces overfitting on small data. It trains on CPU or GPU for classification, regression and ranking, takes text and embedding features, and exports models to ONNX, CoreML, PMML and several runtime languages.
Openness
5 high confidence- license
- Apache-2.0(OSI)
- source
- public(github.com/catboost/catboost)
- core features withheld
- no — Yandex publishes it as open source
Apache-2.0, built from the public repository, and published by Yandex as the owner. Yandex's pages present it as open source for anyone and mention no paid edition or hosted tier.
- https://catboost.ai/ recorded 2026-09-27
catboost.ai: "developed by Yandex researchers and engineers ... It is in open-source and can be used by anyone."; no pricing, cloud or enterprise tier.
- https://raw.githubusercontent.com/catboost/catboost/master/AUTHORS recorded 2026-09-27
AUTHORS: "source code of «CatBoost» published and distributed by YANDEX LLC as the owner".
- https://raw.githubusercontent.com/catboost/catboost/master/LICENSE recorded 2026-09-27
LICENSE opens "Copyright 2017-2026 YANDEX LLC" followed by the Apache License, Version 2.0.
- https://ungh.cc/repos/catboost/catboost/files/master recorded 2026-09-27
Repository tree of 29,822 paths with no ee, enterprise, commercial or pro directory; the only other LICENSE files are third-party ones under contrib/libs.
Adoption
4 high confidenceMeasured on monthly PyPI downloads of the catboost package; the R package and the JVM and C++ appliers are not counted.
- https://pypistats.org/api/packages/catboost/recent recorded 2026-09-27
last_month 5,004,823 downloads of catboost
Capability
2 high confidenceCatBoost is one learner, boosted trees, like XGBoost and LightGBM. What sets it apart is how it treats categorical and text columns without manual encoding and how many places a trained model can be deployed. It does not compare or choose among other kinds of models.
- https://catboost.ai/docs/en/concepts/python-reference_catboost recorded 2026-09-27
Python package reference lists CatBoost, CatBoostClassifier, CatBoostRegressor and CatBoostRanker.
- https://catboost.ai/en/docs/ recorded 2026-09-27
Docs landing: training on GPU, categorical, text and embedding features, applying models from C/C++, Java, Node.js, Rust and ClickHouse, and export to CoreML, JSON, ONNX and PMML.
Verified 2026-09-27