The AI Museum
AI Analytics1,072 recordsRecounted 23 September 2026

Models in 2021

Weightsopenclosedunknownfrontier only
Year20262025202420232022202120202019201820172016
1 filter active, showing 79 of 1,072 recordsClear all

79

Models

39 organisations

52%

Open weights

41 open, 38 closed

8

Frontier

By Epoch's flag

1.6T

Largest traceable

Switch

2021

Span

48 with a traceable size

Releases

by month
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sept
Oct
Nov
Dec

Organisation

Country

Domain

Records

newest 50 of 79
Models in 2021, newest first. Each row links to the source Epoch AI records for it.
ModelOrganisationReleasedParametersWeights
ContrieverMeta AI2021-12110M *Open
ERNIE 3.0 TitanBaidu2021-12260BClosed
ERNIE-ViLGBaidu2021-1210B *Closed
GLaMGoogle2021-121.2TClosed
Gopher (280B)DeepMind2021-12280BClosed
LDM-1.45BHeidelberg University2021-121BOpen
LongT5Google Research2021-123BOpen
Student of GamesDeepMind2021-12—Closed
T-NLRv5 XXLMicrosoft2021-125BClosed
XGLM-7.5BMeta AI2021-128BOpen
AliceMind-MMUAlibaba2021-11—Closed
BASIC-LGoogle2021-113B *Closed
CodeT5-baseSalesforce2021-11220M *Open
FlorenceMicrosoft2021-11893MClosed
Masked Autoencoders ViT-HFacebook AI Research2021-11632M *Open
NÜWAMicrosoft Research2021-11870MClosed
Projected GANHeidelberg University2021-11—Open
Swin Transformer V2 (SwinV2-G)Microsoft Research Asia2021-113BOpen
ViT-G/14 (LiT)Google Research2021-113BClosed
AlphaFold-MultimerGoogle DeepMind2021-10—Open
base LM+GNN+kNNShannon.AI2021-10274MOpen
EfficientZeroTsinghua University2021-10—Closed
EveHarvard Medical School2021-1015M *Closed
Megatron-Turing NLG 530BMicrosoft2021-10530BClosed
S4Stanford University2021-10249M *Open
Yuan 1.0Inspur2021-10246BClosed
FLAN 137BGoogle Research2021-09137BClosed
HyperCLOVA 204BNAVER2021-09204B *Closed
MEBMicrosoft2021-09135BClosed
PermuteFormerPeking University2021-09150M *Closed
PLATO-XLBaidu2021-0911BOpen
TrOCRBeihang University2021-09558MOpen
Turing ULRv5Microsoft2021-092BClosed
6-Act TetherFacebook AI Research2021-085MOpen
DNABERTNortheastern University2021-08110MOpen
Jurassic-1-JumboAI21 Labs2021-08178BClosed
W2v-BERTGoogle Brain2021-081BClosed
XLMR-XXLFacebook AI Research2021-0811BOpen
YOLOX-XMegvii Inc2021-0899M *Open
Zidong TaichuChinese Academy of Sciences2021-083BOpen
CodexOpenAI2021-0712B *Closed
ERNIE 3.0Baidu2021-0710BOpen
GOATDeepMind2021-073M *Closed
HuBERTFacebook AI Research2021-071B *Open
SEERFacebook AI Research2021-071BOpen
Adaptive Input Transformer + RDMicrosoft Research Asia2021-06247M *Closed
ALIGNGoogle Research2021-06820MClosed
CoAtNetGoogle2021-062BClosed
DeBERTaMicrosoft2021-062BOpen
Denoising Diffusion Probabilistic Models (LSUN Bedroom)University of California (UC) Berkeley2021-06256MOpen

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