The AI Museum
AI Analytics1,072 recordsRecounted 23 September 2026

Models from Google DeepMind

Weightsopenclosedunknownfrontier only
Year20262025202420232022202120202019201820172016
1 filter active, showing 47 of 1,072 recordsClear all

47

Models

1 organisations

16%

Open weights

7 open, 36 closed

2

Frontier

By Epoch's flag

55B

Largest traceable

RT-2

2015–2026

Span

10 with a traceable size

Releases

by year
201520162017201820192020202120222023202420252026

Organisation

Country

Domain

Records

47
Models from Google DeepMind, newest first. Each row links to the source Epoch AI records for it.
ModelOrganisationReleasedParametersWeights
Gemini 3.8 FlashGoogle DeepMind2026-09—Closed
WeatherNext 3Google DeepMind2026-09—Closed
Gemini 3.5 FlashGoogle DeepMind2026-05—Closed
Gemini Flash 3.1 TTSGoogle DeepMind2026-04—Closed
Gemini 3.0 Flash-liteGoogle DeepMind2026-03—Closed
SIMA 2Google DeepMind2025-12—Closed
Gemini 3 Deep ThinkGoogle DeepMind2025-11—Closed
Gemini 3 ProGoogle DeepMind2025-11—Closed
Gemini 3 Pro Image (Nano Banana Pro)Google DeepMind2025-11—Closed
Veo 3.1Google DeepMind2025-10—Closed
Gemini Robotics-ER 1.5Google DeepMind2025-09—Closed
Gemini EmbeddingGoogle DeepMind2025-07—Closed
FGNGoogle DeepMind2025-06720MClosed
Gemini 2.5 Pro (Jun 2025)Google DeepMind2025-06—Closed
Gemini 2.5 Pro (May 2025)Google DeepMind2025-05—Closed
Veo 3Google DeepMind2025-05—Closed
Gemini 2.5 Pro (Mar 2025)Google DeepMind2025-03—Closed
Gemini 2.0 ProGoogle DeepMind2024-12—Closed
Veo 2Google DeepMind2024-12—Closed
Gemini-Exp-1114Google DeepMind2024-11—Closed
AlphaProteoGoogle DeepMind2024-09—Closed
Table Tennis AgentGoogle DeepMind2024-08185,000 *Closed
AlphaFold 3Google DeepMind2024-05—Open
GenCastGoogle DeepMind2024-05—Open
Gemini 1.5 ProGoogle DeepMind2024-02—Closed
AlphaGeometryGoogle DeepMind2024-01151MOpen
FunSearchGoogle DeepMind2023-1215B *Open
Gemini 1.0 ProGoogle DeepMind2023-12—Closed
Gemini 1.0 UltraGoogle DeepMind2023-12—Closed
Gemini Nano-1Google DeepMind2023-122BClosed
Gemini Nano-2Google DeepMind2023-123BClosed
GNoME for crystal discoveryGoogle DeepMind2023-1116M *Closed
GraphCastGoogle DeepMind2023-11—Open
RT-TrajectoryGoogle DeepMind2023-11—Closed
PaLI-3Google DeepMind2023-105BClosed
RT-2-XGoogle DeepMind2023-1055BClosed
AlphaMissenseGoogle DeepMind2023-0993M *Closed
RT-2Google DeepMind2023-0755BClosed
Agile Soccer RobotGoogle DeepMind2023-04—Closed
SigLIP 400MGoogle DeepMind2023-03400MOpen
AlphaFold-MultimerGoogle DeepMind2021-10—Open
WaveNetGoogle DeepMind2016-09—Not recorded
PixelCNNGoogle DeepMind2016-06—Not recorded
Dueling DQNGoogle DeepMind2016-042MClosed
Double DQNGoogle DeepMind2016-032MClosed
Advantage LearningGoogle DeepMind2015-12—Not recorded
Deep Deterministic Policy GradientsGoogle DeepMind2015-09—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