OPT-175B Research Access Announcement
Meta AI announced on May 3, 2022 that it was sharing Open Pretrained Transformer (OPT-175B), a 175-billion-parameter language model, with the AI research community under a research-access program.
Historical classification
This verified record belongs to LINEAiGE's artificial-intelligence history layer beginning in 1956.
Why it matters
A documented large-language-model access milestone in which Meta paired a 175-billion-parameter model with research access and supporting research artifacts intended to broaden independent study of large language models.
People, organizations & technologies
ORGANIZATIONS
Meta AI
TECHNOLOGIES
OPT-175B
Evidence & verification
STATE
verified
CONFIDENCE
high
LAST VERIFIED
2026-09-09T22:04:25+08:00
Meta AI's first-party May 3, 2022 announcement states that it was sharing OPT-175B, a 175-billion-parameter language model trained on publicly available datasets, for research access. The associated paper's arXiv record was submitted May 2, 2022 and describes OPT as a suite of decoder-only pretrained transformers from 125M to 175B parameters, with code and a training logbook released and model access intended for interested researchers. This record is intentionally classified as a research-access announcement, not a general-public product launch, unrestricted open-weight release, API launch, deployment date, or causal ancestor of later models. No geography, ancestry, influence, causality, or relationships are inferred.
Sources
Democratizing access to large-scale language models with OPT-175B
Meta AI
PUBLISHED
2022-05-03
OPT: Open Pre-trained Transformer Language Models
arXiv
PUBLISHED
2022-05-02
Trace the chronology
Record identity
LINEAiGE IDopt-175b-2022