FineWeb

license

ODC-BY

gated

false

dataset_card

present

format

parquet

Permissive ODC-BY license with ungated download and a full dataset card.

• https://huggingface.co/api/datasets/HuggingFaceFW/fineweb recorded 2026-09-29

  Hub API record for HuggingFaceFW/fineweb listing the tags "task_categories:text-generation" and "language:en" and "license:odc-by" for the dataset, so the license is ODC-By and the gated field is false.

• https://huggingface.co/datasets/HuggingFaceFW/fineweb recorded 2026-09-29

  Hugging Face dataset page for HuggingFaceFW/fineweb rendering the card, with the license "Open Data Commons Attribution License (ODC-By) v1.0".

• https://huggingface.co/datasets/HuggingFaceFW/fineweb/raw/main/README.md recorded 2026-09-29

  Raw dataset card README for HuggingFaceFW/fineweb, which lists "- **License:** Open Data Commons Attribution License (ODC-By) v1.0" as its license.

Adoption is measured by Hugging Face downloads of the dataset over the trailing month, the only usage channel on record for FineWeb.

• https://huggingface.co/api/datasets/HuggingFaceFW/fineweb recorded 2026-08-13

  418,578 downloads in the trailing 30 days for HuggingFaceFW/fineweb

FineWeb outperforms C4, RefinedWeb, RedPajama2, Dolma, The Pile and SlimPajama in the paper's matched 1.71B-parameter ablations. FineWeb-Edu, its educational-quality subset, scores well above it on MMLU, ARC and OpenBookQA.

• https://arxiv.org/abs/2406.17557 recorded 2026-09-29

  arXiv abstract for the FineWeb paper describes "a 15-trillion token dataset derived from 96 Common Crawl snapshots that produces better-performing LLMs than other open pretraining datasets".

• https://arxiv.org/html/2406.17557 recorded 2026-09-29

  FineWeb paper HTML compares FineWeb in Fig. 10 against datasets including "the deduplicated variant of RedPajama2" and "Colossal-OSCAR (850B tokens)", reports that "FineWeb shows strong performance and FineWeb-Edu (detailed below) outperforms all other open datasets we compared on our aggregate group of tasks", and says FineWeb-Edu "outperforms all openly accessible web-based datasets on a number of reasoning- and knowledge-intensive benchmarks such as MMLU, ARC, and OpenBookQA by a significant margin".