Organization
DeepMind
Tracked models
79
Frontier models
6
Largest training run
6.3 x 10^23 FLOP
79 tracked AI models list DeepMind as a developer. 6 of them are frontier models — among the largest training runs of their moment. The largest is Gopher (280B), at 6.3 x 10^23 FLOP (December 2021). The most recent tracked release is from May 2025.
Models (79)
Gopher (280B)6.3 x 10^23 FLOPChinchilla5.8 x 10^23 FLOPAlphaCode2.4 x 10^23 FLOPFlamingo2.2 x 10^23 FLOPAlphaStar1.1 x 10^23 FLOPStudent of Games3.7 x 10^22 FLOPGOAT2.4 x 10^22 FLOPRETRO-7B1.7 x 10^22 FLOP$\infty$-former (SM)1.2 x 10^22 FLOPDistilled Grandmaster10^22 FLOPSILC-S* (86M)10^22 FLOPDCTransformer (ImageNet)4.4 x 10^21 FLOPAlphaFold-Multimer4.3 x 10^21 FLOPGato4 x 10^21 FLOPAlphaFold 23 x 10^21 FLOPAdaptive Agent2.8 x 10^21 FLOPJEST-L++2 x 10^21 FLOPEMDR1.9 x 10^21 FLOPAlphaGo Lee1.9 x 10^21 FLOPBigGAN-deep 512x5121.8 x 10^21 FLOPAlphaTensor7.1 x 10^20 FLOPAlphaGo Zero6.5 x 10^20 FLOPOcto-Base5.8 x 10^20 FLOPAlphaGo Fan3.8 x 10^20 FLOPAlphaGo Master3.4 x 10^20 FLOPDreamerV32.2 x 10^20 FLOPIMPALA1.7 x 10^20 FLOPAlphaZero1.1 x 10^20 FLOPLong-range sequence Compressive Transformers10^20 FLOPAlphaFold10^20 FLOPMuZero4.8 x 10^19 FLOPFTW (For The Win)3.5 x 10^19 FLOPLSTM (Hebbian, Cache, MbPA)3.3 x 10^19 FLOPT2R + Random Init2.7 x 10^19 FLOPT2R + Pretrain1.4 x 10^19 FLOPT2R 75% + Pretrain (WT-103)1.4 x 10^19 FLOPRFA-GATE-Gaussian-Stateful Big7.1 x 10^18 FLOPHanabi 4 player4.3 x 10^18 FLOPDARTS3.2 x 10^17 FLOPMogrifier RLSTM (WT2)1.4 x 10^17 FLOPMogrifier RLSTM (PTB)7.1 x 10^16 FLOPDQN2.8 x 10^15 FLOPAlphaEvolveMay 2025SO3LROctober 2024EnzymeFlowOctober 2024Octo-SmallMay 2024SILC-SOctober 2023Med-PaLMJuly 2023Med-PaLM 2May 2023E-SPAFebruary 2023DeepNashDecember 2022SparrowSeptember 2022MuZero VP9February 2022Transformer-XL + RelationLMJanuary 2022SPALM + RelationLMJanuary 2022Gopher (7.1B)December 2021SPALM + kNNApril 2021Agent57March 2020Perceiver IO (optical flow)February 2020Mogrifier (d2, MC) + dynamic evalSeptember 2019Mogrifier (d2, MoS2, MC) + dynamic evalSeptember 2019CPC v2May 2019Neuro-Symbolic Concept LearnerApril 2019SPN (ImageNet 128)December 2018SPN (CelebA HQ)December 2018DARTS (second order) (PTB)June 2018Relational Memory CoreJune 20182-layer skip-LSTM + dropout tuning (WT2)May 20182-layer skip-LSTM + dropout tuning (PTB)May 2018Multipop Adaptive Continuous Stack (WT2)February 2018Multipop Adaptive Continuous Stack (PTB)February 2018VQ-VAENovember 2017Rainbow DQNOctober 2017AWD-LSTMJuly 2017NoisyNet-DuelingJune 2017Reading Twice for NLUJune 2017Inflated 3D ConvNetJune 2017Elastic weight consolidationDecember 2016RNN-WERJune 2014
SourceEpoch AI, 'AI Models'. Published online at epoch.ai. Retrieved 2026-07-29 from https://epoch.ai/data/ai-models. Licensed under CC BY 4.0.
Other organizations