Evidence-backed European AI register
BigScience
Reviewed identity, roles, model listings, pricing and evidence for BigScience.
Dataset compiled 2026-08-27 · evidence over inference
Reviewed identity
- Headquarters scope
- EU headquarters
- Headquarters record
- Paris (coordinated from France; international collaboration)
- Legal entity
- open research collaboration (no legal entity); coordinated by Hugging Face with CNRS/GENCI compute
- Ownership
- non-profit / open collaboration
- Founded
- 2021
- Confidence
- high
- Last verified
- 2026-08-22
Evidence boundaries
Organisation listings identify a reviewed canonical record. Model listings show an association in the source register; they do not by themselves establish model creation, current API availability, hosting location or infrastructure ownership.
Model listings
- Type
- text
- Parameters
- 176B
- Context
- 2,048 tokens
- Languages
- 46 natural languages + 13 programming languages
- Licence
- custom (BigScience RAIL License v1.0)
- Release
- 2022-07
Profile note
Dormant but foundational: the 2021–22 BigScience workshop proved that European public HPC (Jean Zay) plus an open international collaboration could train a frontier-scale LLM, and set the transparency norms (ROOTS datasheet, RAIL licence) that later French labs inherited. No new releases; recorded as legacy.
Record sources
- https://huggingface.co/bigscience/bloom
- https://arxiv.org/html/2211.05100
- https://www.genci.fr/en/news/delivery-largest-open-science-multilingual-language-model-ever-produced
- https://www.cnrs.fr/en/press/release-largest-trained-open-science-multilingual-language-model-ever
- https://huggingface.co/bigscience