The AI Museum
AI Analytics1,072 recordsRecounted 23 September 2026

Open-weight language models from United States of America

Weightsopenclosedunknownfrontier only
Year20262025202420232022202120202019201820172016
3 filters active, showing 81 of 1,072 recordsClear all

81

Models

29 organisations

100%

Open weights

81 open, 0 closed

9

Frontier

By Epoch's flag

1.6T

Largest traceable

Switch

2016–2026

Span

53 with a traceable size

Releases

by year
20162017201820192020202120222023202420252026

Organisation

Country

Domain

Records

newest 50 of 81
Open-weight language models from United States of America, newest first. Each row links to the source Epoch AI records for it.
ModelOrganisationReleasedParametersWeights
InklingThinking Machines2026-07975B *Open
Nemotron 3 UltraNVIDIA2026-06550B *Open
Nemotron 3 SuperNVIDIA2026-03120B *Open
Nemotron 3-Nano-30B-A3BNVIDIA2025-1232B *Open
Olmo 3Allen Institute for AI2025-1132BOpen
gpt-oss-120bOpenAI2025-08117BOpen
gpt-oss-20bOpenAI2025-0821BOpen
Llama 3.3 70BMeta AI2024-1270BOpen
Llama-3.1-Nemotron-70B-InstructNVIDIA2024-10—Open
Llama 3.1-405BMeta AI2024-07405BOpen
Nemotron-4 340BNVIDIA2024-06340BOpen
Llama 3-70BMeta AI2024-0470BOpen
DBRXDatabricks2024-03132BOpen
FunSearchGoogle DeepMind2023-1215B *Open
Llama GuardMeta AI2023-127BOpen
Grok-1xAI2023-11314B *Open
Nemotron-3-8BNVIDIA2023-118BOpen
FinGPT-13BUniversity of California Los Angeles (UCLA)2023-1013B *Open
JaisCerebras Systems2023-0813BOpen
Llama 2-70BMeta AI2023-0770BOpen
Llama 2-7BMeta AI2023-077BOpen
CodeT5+Salesforce2023-0516BOpen
CoEdiT-xxlUniversity of Minnesota2023-0511B *Open
StarCoderHugging Face2023-0516BOpen
Incoder-6.7BFacebook AI Research2023-047BOpen
LLaMA-65BMeta AI2023-0265BOpen
Hybrid H3-2.7BStanford University2022-123B *Open
GalacticaMeta AI2022-11120B *Open
mT0-13BHugging Face2022-1113BOpen
BLOOM-176BHugging Face2022-07176BOpen
CodeT5-largeSalesforce2022-07770M *Open
NLLBMeta AI2022-0755BOpen
GPT-2 Medium (FlashAttention)Stanford University2022-05355MOpen
OPT-175BMeta AI2022-05175BOpen
SimCSEPrinceton University2022-05—Open
UL2Google Research2022-0520BOpen
GPT-NeoX-20BEleutherAI2022-0220BOpen
PolyCoderCarnegie Mellon University (CMU)2022-023B *Open
data2vec (language)Meta AI2022-01705M *Open
ContrieverMeta AI2021-12110M *Open
LongT5Google Research2021-123BOpen
XGLM-7.5BMeta AI2021-128BOpen
CodeT5-baseSalesforce2021-11220M *Open
S4Stanford University2021-10249M *Open
XLMR-XXLFacebook AI Research2021-0811BOpen
DeBERTaMicrosoft2021-062BOpen
ByT5-XXLGoogle2021-0513B *Open
Generative BSTFacebook AI Research2021-039BOpen
UnicornAllen Institute for AI2021-0311BOpen
SRU++ LargeASAPP2021-02234MOpen

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