XGBoost
DMLCOptimized gradient-boosted decision tree library for classification, regression, ranking and survival analysis, running on a single machine, on GPUs, and distributed across Kubernetes, Spark, Dask and Ray. Bindings cover Python, R, the JVM and several other languages. It began as a University of Washington research project and is governed by a project management committee.
Openness
5 high confidence- license
- Apache-2.0(OSI)
- source
- public(github.com/dmlc/xgboost)
- core features withheld
- no — committee-governed
Apache-2.0, built from the public repository, and run by a project management committee that takes sponsorship for its CI costs. No company sells an edition of it.
- https://raw.githubusercontent.com/dmlc/xgboost/HEAD/LICENSE recorded 2026-09-26
LICENSE is the Apache License, Version 2.0, "Copyright (c) 2019 by Contributors".
- https://raw.githubusercontent.com/dmlc/xgboost/HEAD/README.md recorded 2026-09-26
README: "XGBoost is an optimized distributed gradient boosting library"; License section "Licensed under an Apache-2 license"; sponsors listed via Open Source Collective; no paid tier mentioned.
- https://ungh.cc/repos/dmlc/xgboost/files/HEAD recorded 2026-09-26
Repository tree of 1,453 paths with no ee, enterprise, commercial or pro directory.
- https://xgboost.readthedocs.io/en/stable/contrib/community.html recorded 2026-09-26
Community guide: "XGBoost adopts the Apache style model and governs by merit." and "The PMC may accept 3rd-party donations and sponsorships that would defray the cost of the CI infrastructure."
Adoption
5 high confidenceMeasured on monthly PyPI downloads of the xgboost package; the R, JVM and other bindings are not counted.
- https://pypistats.org/api/packages/xgboost/recent recorded 2026-09-26
last_month 29,819,156 downloads of xgboost
Capability
2 high confidenceXGBoost does one kind of model, boosted trees, and does it very well, at scale and on GPUs. Anything else - another learner, the comparison between them, the pipeline around them - comes from a toolkit like scikit-learn, whose estimator interface it implements.
- https://xgboost.readthedocs.io/en/stable/ recorded 2026-09-26
Docs landing tutorials: Introduction to Boosted Trees, Learning to Rank, DART, Monotonic Constraints, Feature Interaction Constraints, Survival Analysis with Accelerated Failure Time, Categorical Data, Multiple Outputs, Random Forests in XGBoost; distributed on Kubernetes, Spark, Dask, PySpark and Ray; GPU support; scikit-learn estimator interface.
Verified 2026-09-26