Tier 5 Fully open / reproducible
What the publisher claims: open source
The one-liner: MIT-licensed gradient boosting with no attachments.
The six questions
Full source and binaries are downloadable.
MIT.
Runs locally or distributed with no proprietary components.
Algorithm, defaults and code are fully public and the training data is the user's own.
A skilled group can rebuild an equivalent implementation from the published algorithm and code.
LightGBM is hosted by Microsoft but runs on a public roadmap issue and a long-standing outside contributor community that merges PRs directly (about 31 non-bot external merges in the latest 75), with community voice in feature prioritisation.
License
MIT OSI-approved — license source ↗
What is actually released
- library source
- documentation
- papers
Why it matters
Confirms the classical-ML openness baseline across independent implementations rather than a single project.
Sources
Derived from the openness dataset behind But Is It Open? — snapshot 2026-11-07 ↗. Scores are qualitative readings of the cited evidence, not verdicts; tier measures openness, not intelligence.