AI Potluck
Back to Gap Map Model components / Data hubs & catalogs

OpenML

OpenML
open source / Overall score: n/a

OpenML is an open platform for sharing machine learning datasets together with the tasks, pipelines and experiment runs built on them. Anyone can upload a dataset from a dataframe; OpenML stores it in a uniform format under an ID and version, computes its data characteristics, and serves it to Python, R and Java clients. It is run as a community project represented by the not-for-profit Open Machine Learning Foundation.

Scored as the hosted platform at openml.org. Its PHP backend is in maintenance mode while a Python rewrite of the server is developed in a separate repository.

Openness

5 high confidence
5.0
license
BSD-3-Clause(OSI)
source
public(github.com/openml/OpenML)
core features withheld
no — community project represented by the not-for-profit Open Machine Learning Foundation

The server behind openml.org is published under BSD-3-Clause, and so is the rewrite that will replace it. OpenML is a community project backed by a not-for-profit foundation, with no paid edition.

Adoption

not assessed

OpenML publishes no total of downloads or users, and its search index refused automated queries. Downloads of the openml client packages would count a client rather than the platform, so no reading was recorded.

Capability

5 high confidence
5.0

Anyone can upload a dataset from a dataframe, and OpenML stores it in one uniform format, computes its characteristics and serves it back by ID to Python, R and Java. Tasks and benchmark suites are defined on top of the datasets, so the catalog also records how each one is used for evaluation.

Verified 2026-10-07