AI Analytics1,072 recordsRecounted 23 September 2026

Unrecorded-weight frontier models

Weightsopen closed unknown frontier only

Year2026 2025 2024 2023 2022 2021 2020 2019 2018 2017 2016

2 filters active, showing 60 of 1,072 recordsClear all

Models

30 organisations

—

Open weights

0 open, 0 closed

Frontier

By Epoch's flag

Largest traceable

Seq2Seq LSTM

1950–2017

Span

31 with a traceable size

Releases

by year, from 2002

2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Records

newest 50 of 60

Unrecorded-weight frontier models, newest first. Each row links to the source Epoch AI records for it.
Model	Organisation	Released	Parameters	Weights
OpenAI TI7 DOTA 1v1	OpenAI	2017-08	—	Not recorded
JFT	Google Research	2017-07	45M	Not recorded
NASv3 (CIFAR-10)	Google Brain	2016-11	37M *	Not recorded
Xception	Google	2016-10	23M	Not recorded
Inception v3	Google	2015-12	24M *	Not recorded
MSRA (C, PReLU)	Microsoft Research	2015-02	87M	Not recorded
SNM-skip	Google	2014-12	62B *	Not recorded
RNNsearch-50*	Jacobs University Bremen	2014-09	—	Not recorded
Seq2Seq LSTM	Google	2014-09	2B	Not recorded
VGG16	University of Oxford	2014-09	138M	Not recorded
VGG19	University of Oxford	2014-09	144M *	Not recorded
SPPNet	Microsoft	2014-06	—	Not recorded
TransE	Universite de Technologie de Compiègne – CNRS	2013-12	942M *	Not recorded
Visualizing CNNs	New York University (NYU)	2013-11	—	Not recorded
DistBelief NNLM	Google	2013-01	—	Not recorded
AlexNet	University of Toronto	2012-09	60M	Not recorded
Unsupervised High-level Feature Learner	Google	2012-07	1B *	Not recorded
Deep Autoencoders	University of Toronto	2011-04	140M	Not recorded
RNN LM	Johns Hopkins University	2010-09	70M *	Not recorded
LCNP LabelMe	University of Bonn	2009-11	14M	Not recorded
LCNP MNIST		2009-11	12M	Not recorded
KN-LM	Google	2007-06	21B *	Not recorded
SB-LM	Google	2007-06	300B *	Not recorded
Hierarchical LM		2005-01	—	Not recorded
LMICA		2004-12	4M	Not recorded
NPLM (AP News)	University of Montreal / Université de Montréal	2003-03	12M	Not recorded
NPLM (Brown)	University of Montreal / Université de Montréal	2003-03	4M	Not recorded
Decision tree (classification)	Mitsubishi Electric Research Labs	2001-12	12,000 *	Not recorded
Neural LM	University of Montreal / Université de Montréal	2000-11	7M	Not recorded
PoE MNIST	University College London (UCL)	2000-11	4M	Not recorded
RECONTRA-categorized		1999-06	66,780 *	Not recorded
LeNet-5	AT&T	1998-11	60,000	Not recorded
LSTM	Technical University of Munich	1997-11	10,504	Not recorded
MUSIC perceptron		1996-06	13,607	Not recorded
NeuroChess		1994-12	72,251 *	Not recorded
Predictive Coding NN	Technical University of Munich	1994-12	206,910	Not recorded
Siamese-TDNN	Bell Laboratories	1993-08	744 *	Not recorded
TD-Gammon	IBM	1992-05	25,000 *	Not recorded
Weight Decay		1991-12	8,386	Not recorded
SexNet compression		1990-10	72,940	Not recorded
NETtalk reimplementation	Oregon State University	1990-06	27,480	Not recorded
Zip CNN	AT&T	1989-12	9,760	Not recorded
Handwritten digit recognition network	AT&T	1989-11	2,578	Not recorded
Invariant image recognition	Complutense University of Madrid	1989-06	—	Not recorded
NetTalk (dictionary)	Princeton University	1987-06	18,629	Not recorded
NetTalk (transcription)	Princeton University	1987-06	18,629	Not recorded
Translation-invariant MLP	Carnegie Mellon University (CMU)	1987-06	816	Not recorded
MLP with back-propagation	University of California San Diego	1986-10	720	Not recorded
Distributed representation NN	Carnegie Mellon University (CMU)	1986-08	432	Not recorded
ASE+ACE	University of Massachusetts Amherst	1983-09	324 *	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

Unrecorded-weight frontier models

Releases

Organisation

Country

Domain

Records