The AI Museum
AI Analytics1,092 recordsRecounted 5 November 2026

Language models in 2026

Weightsopenclosedunknownfrontier only
Year20262025202420232022202120202019201820172016
2 filters active, showing 71 of 1,092 recordsClear all

71

Models

27 organisations

42%

Open weights

29 open, 40 closed

0

Frontier

By Epoch's flag

2.8T

Largest traceable

MAI-Base-2

2026

Span

17 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 71
Language models in 2026, newest first. Each row links to the source Epoch AI records for it.
ModelOrganisationReleasedParametersWeights
Gemini 4 CarbonGoogle DeepMind2026-10—Closed
Grok Voice 3.0xAI2026-10—Closed
MAI-Base-2Microsoft2026-102.8TClosed
Mistral Large 4Mistral AI2026-101.1TOpen
Muse WonderMeta AI2026-10—Closed
Muse Wonder FullMeta AI2026-10—Closed
Claude Fable 5.1Anthropic2026-09—Closed
Claude Opus 5.5Anthropic2026-09—Closed
Claude Sonnet 5.5Anthropic2026-09—Closed
DeepSeek V4.1 FlashDeepSeek2026-09552B *Open
Gemini 3.8 FlashGoogle DeepMind2026-09—Closed
Gemini 4 ArgonGoogle DeepMind2026-09—Not recorded
GPT-6 Luna DecisionsOpenAI2026-09—Closed
Grok 4.7xAI2026-092B *Closed
Mercury 2.5Inception Labs2026-09—Closed
Muse Spark 1.3Meta AI2026-09—Closed
OpenEvidence DarwinOpenEvidence2026-09—Closed
Step 5 PreviewStepFun2026-09600B *Closed
GLM-5.3Z.ai (Zhipu AI)2026-08744BOpen
GLM-5.3-FlashZ.ai (Zhipu AI)2026-08320B *Open
Motif-3Motif Technologies2026-08314BOpen
Muse Spark 1.2Meta AI2026-08—Closed
Qwen3.8-2.4T-A95BAlibaba2026-082.4TOpen
SeedRealtimeByteDance2026-08—Closed
Tencent Hy4 previewTencent2026-08770B *Open
A.X K2SK Telecom2026-07688BOpen
Claude Opus 5Anthropic2026-07—Closed
InklingThinking Machines2026-07975B *Open
K-EXAONE 2.0LG AI Research2026-07750BOpen
Kimi K3Moonshot2026-072.8T *Open
Qwen 3.8 MaxAlibaba2026-072.4TClosed
Tencent Hy3Tencent2026-07295BOpen
Claude Fable 5Anthropic2026-06—Closed
Claude Mythos 5Anthropic2026-06—Closed
LongCat-2.0Meituan Inc2026-061.6T *Open
MAI-Code-1-FlashMicrosoft2026-06—Closed
MiniMax-M3MiniMax2026-06428BOpen
Nemotron 3 UltraNvidia2026-06550B *Open
Qwen3.7-PlusAlibaba2026-06—Closed
Solar Open2 250BUpstage2026-06250B *Open
Claude Opus 4.8Anthropic2026-05—Closed
Composer 2.5Cursor2026-051.0T *Closed
Gemini 3.5 FlashGoogle DeepMind2026-05—Closed
TML-Interaction-SmallThinking Machines2026-05276BClosed
Claude Mythos PreviewAnthropic2026-04—Not recorded
Claude Opus 4.7Anthropic2026-04—Closed
DeepSeek-V4-FlashDeepSeek2026-04284B *Open
DeepSeek-V4-ProDeepSeek2026-041.6T *Open
EXAONE 4.5LG AI Research2026-0433BOpen
GLM-5.1Z.ai (Zhipu AI)2026-04754BOpen

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