MBPP

license

CC-BY-4.0(open data license)

redistributability

full(downloadable, ungated)

datasheet

yes(extensive dataset card)

size

~974 full / 427 sanitized problems w/ test_list

splits

train/test/validation/prompt public

Fully open: CC-BY-4.0, ungated, downloadable, with an extensive dataset card. Clean open-data case; the standard companion to HumanEval for basic Python code-gen eval.

• https://huggingface.co/datasets/google-research-datasets/mbpp recorded 2026-09-29

  The dataset page carries the tag "license:cc-by-4.0", the embedded repo state reads "gated":false, and the card opens with "Dataset Card for Mostly Basic Python Problems (mbpp)" and a "Licensing Information" section; searched the body for non-commercial and holdout clauses: none.

Adoption is measured through Hugging Face downloads of google-research-datasets/mbpp. MBPP is a standard basic-Python code-generation benchmark, routinely paired with HumanEval in code-model release reports, and this category counts dataset downloads on a far smaller scale than it does for software or models.

• https://huggingface.co/api/datasets/google-research-datasets/mbpp recorded 2026-08-12

  208,317 downloads in the trailing 30 days for google-research-datasets/mbpp

A dataset is not 'capable', so capability is left unscored. MBPP is largely saturated by frontier models, but that is a quality caveat rather than a capability score.

• https://huggingface.co/datasets/google-research-datasets/mbpp recorded 2026-09-29

  The dataset page describes "around 1,000 crowd-sourced Python programming problems, designed to be solvable by entry level programmers" with "a task description, code solution and 3 automated test cases"; searched the body for accuracy and SOTA claims: none.