The AI Museum
AI Analytics1,072 recordsRecounted 23 September 2026

Unrecorded-weight language models

Weightsopenclosedunknownfrontier only
Year20262025202420232022202120202019201820172016
2 filters active, showing 54 of 1,072 recordsClear all

54

Models

31 organisations

—

Open weights

0 open, 0 closed

16

Frontier

By Epoch's flag

2B

Largest traceable

Seq2Seq LSTM

1959–2026

Span

10 with a traceable size

Releases

by year, from 2011
2011201220132014201520162017201820192020202120222023202420252026

Organisation

Country

Domain

Records

newest 50 of 54
Unrecorded-weight language models, newest first. Each row links to the source Epoch AI records for it.
ModelOrganisationReleasedParametersWeights
Claude Mythos PreviewAnthropic2026-04—Not recorded
GPipe (Transformer)Google2018-116B *Not recorded
Chinese - English translationMicrosoft2018-03—Not recorded
ELMoUniversity of Washington2018-0294M *Not recorded
PhraseCondCarnegie Mellon University (CMU)2017-10—Not recorded
S-NormUniversity of Washington2017-10—Not recorded
ConvS2S (ensemble of 8 models)Meta AI2017-07—Not recorded
GSMPeking University2017-07—Not recorded
Reading Twice for NLUDeepMind2017-06—Not recorded
Mnemonic ReaderFudan University2017-05—Not recorded
Character-enriched word2vecFacebook AI Research2016-07—Not recorded
Named Entity Recognition modelCarnegie Mellon University (CMU)2016-03—Not recorded
BPEUniversity of Edinburgh2015-08—Not recorded
SNM-skipGoogle2014-1262B *Not recorded
RNNsearch-50*Jacobs University Bremen2014-09—Not recorded
Seq2Seq LSTMGoogle2014-092BNot recorded
AdaRNNBeihang University2014-0613,040Not recorded
GRUsUniversity of Montreal / Université de Montréal2014-06—Not recorded
Paragraph VectorGoogle2014-0532MNot recorded
GloVe (32B)Stanford University2014-01120M *Not recorded
GloVe (6B)Stanford University2014-01120M *Not recorded
RNN for 1B wordsGoogle2013-1220B *Not recorded
TransEUniversite de Technologie de Compiègne – CNRS2013-12942M *Not recorded
RCTMUniversity of Oxford2013-10—Not recorded
Word2Vec (large)Google2013-10692M *Not recorded
Word2Vec (small)Google2013-10208M *Not recorded
DistBelief NNLMGoogle2013-01—Not recorded
LSTM LMRWTH Aachen University2012-09103M *Not recorded
Context-dependent RNNMicrosoft Research2012-07—Not recorded
NLP from scratchNEC Laboratories2011-115M *Not recorded
Adaptive SubgradTechnion - Israel Institute of Technology2011-10—Not recorded
Cross-Lingual POS TaggerCarnegie Mellon University (CMU)2011-06—Not recorded
Vector Space ModelStanford University2011-06255,000Not recorded
RNN LMJohns Hopkins University2010-0970M *Not recorded
HLBLUniversity of Toronto2008-122MNot recorded
Deep Multitask NLP NetworkNEC Laboratories2008-072M *Not recorded
KN-LMGoogle2007-0621B *Not recorded
SB-LMGoogle2007-06300B *Not recorded
λ-WASPUT Austin2007-06—Not recorded
Hierarchical LM2005-01—Not recorded
NPLM (AP News)University of Montreal / Université de Montréal2003-0312MNot recorded
NPLM (Brown)University of Montreal / Université de Montréal2003-034MNot recorded
LDAStanford University2003-02—Not recorded
Thumbs Up?Cornell University2002-05—Not recorded
Immediate triheadBrown University2001-07—Not recorded
Neural LMUniversity of Montreal / Université de Montréal2000-117MNot recorded
IBM Model 4University of Southern California1999-07—Not recorded
RECONTRA-categorized1999-0666,780 *Not recorded
RECONTRA-uncategorized1999-06112,000 *Not recorded
LSTM with forget gatesIDSIA1999-01276 *Not recorded

An asterisk marks a parameter figure the source does not consider traceable to the people who built the model. That judgement is the source’s and it has false negatives: Kimi K3 is marked speculative at 2.8T while Moonshot’s own model card states the figure. Why that distinction carries the whole exhibit.

Counts are of models Epoch AI records as notable, not of every model released. The most recent month is always incomplete, because a model enters the dataset when it is reviewed rather than when it ships.

Epoch AI, 'Data on AI Models'. Published online at epoch.ai. Retrieved from 'https://epoch.ai/data/ai-models-documentation' Used under CC BY 4.0.

How this museum is made