MultiPL-E

data

downloadable(HF parquet, 47 subsets)

license

BSD-3-Clause-with-ML-Restriction(the nuprl/MultiPL-E repository LICENSE, whose clause 4 forbids using the contents as training data for any machine learning model

datasheet

present(card documents prompts/doctests/tests/stop-tokens)

redistributable

yes(HF)

MultiPL-E's GitHub repository ships under a modified BSD 3-Clause license whose fourth clause forbids using the contents as training data for any machine-learning model; where the Hugging Face copy's plain MIT tag disagrees, the repository's own license governs. The dataset downloads freely and may be used commercially, but not to train a model. A license that allows commercial use while barring one class of use is bounded, so its tier is use_bounded.

• https://github.com/nuprl/MultiPL-E/blob/main/LICENSE recorded 2026-09-28

  LICENSE headed "BSD 3-Clause License with Machine Learning Restriction"; clause 4 reads "The contents of this repository may not be used as training data for any machine learning model, including but not limited to neural networks."

• https://huggingface.co/datasets/nuprl/MultiPL-E recorded 2026-09-29

  The Hub page shows the tag "license:mit", the state "gated":false, and the card text "MultiPL-E is a dataset for evaluating large language models for code".

• https://raw.githubusercontent.com/nuprl/MultiPL-E/main/LICENSE recorded 2026-09-29

  The raw LICENSE is headed "BSD 3-Clause License with Machine Learning Restriction" and its clause 4 says "may not be used as training data for any".

Adoption is measured by Hugging Face downloads of nuprl/MultiPL-E, the dataset's only declared artifact.

• https://huggingface.co/api/datasets/nuprl/MultiPL-E recorded 2026-08-13

  downloads 33,673 in the trailing 30 days on the declared dataset artifact

MultiPL-E is a dataset and translation framework rather than a system with capabilities to measure, so capability is left unscored.

• https://huggingface.co/datasets/nuprl/MultiPL-E recorded 2026-09-29

  Hugging Face dataset card stating "MultiPL-E is a dataset for evaluating large language models for code" across 22 programming languages, i.e. a static set of translated prompts and tests; searched the body for SOTA, throughput and pass@1: none.