Country
Germany
Tracked models
109
Frontier models
4
Largest training run
4.4 x 10^23 FLOP
109 tracked AI models have at least one contributing organization based in Germany. 4 of them are frontier models — among the largest training runs of their moment. The largest is Pharia-1-LLM-7B, at 4.4 x 10^23 FLOP (August 2024). The most recent tracked release is from November 2025.
It hosts 11 tracked GPU clusters.
Models (80+)
Pharia-1-LLM-7B4.4 x 10^23 FLOPLuminous-supreme3.5 x 10^23 FLOPTeuken 7B2.1 x 10^23 FLOPLuminous-extended10^23 FLOPStarCoder8.5 x 10^22 FLOPCLIP ViT-H/14 - LAION-2B7.8 x 10^22 FLOPProtT5-XXL7.4 x 10^22 FLOPStable Diffusion 1.45 x 10^22 FLOPStable Diffusion (LDM-KL-8-G)5 x 10^22 FLOPProtBERT-BFD3.9 x 10^22 FLOPProtT5-XXL-BFD3.7 x 10^22 FLOPLuminous-base3.2 x 10^22 FLOPPrithvi-EO-2.0 600M2 x 10^22 FLOPProtT5-XL-U501.9 x 10^22 FLOPGPT-J-6B1.5 x 10^22 FLOPSILC-S* (86M)10^22 FLOPWuerstchen8.3 x 10^21 FLOPNucleotide Transformer8.1 x 10^21 FLOPProteinStructureTransformer7.6 x 10^21 FLOPProtBERT-UniRef7.3 x 10^21 FLOPPrithvi-EO-2.0 300M7.1 x 10^21 FLOPAnkh_large6.5 x 10^21 FLOPZymCTRL5 x 10^21 FLOPProtGPT24.1 x 10^21 FLOPTerraMind3.1 x 10^21 FLOPProstT53.1 x 10^21 FLOPAnkh_base2.6 x 10^21 FLOPGBERT-Large2.2 x 10^21 FLOPTiTok-L1.7 x 10^21 FLOPGerman ELECTRA Large1.4 x 10^21 FLOPGenSLM1.4 x 10^21 FLOPStyleGAN-XL1.3 x 10^21 FLOPMACE-MP-08.8 x 10^20 FLOPGemNet-OC5.4 x 10^20 FLOPSeqVec4.1 x 10^19 FLOPProjected GAN1.1 x 10^19 FLOPTrellisNet2.8 x 10^18 FLOPEGRU (WT2)2.2 x 10^18 FLOPMachine learning a model for RNA structure prediction1.8 x 10^18 FLOPRNNsearch-50*1.6 x 10^18 FLOPGemNet-T (OC20)7.3 x 10^17 FLOPUDSMProt6.4 x 10^17 FLOPDecoupled weight decay regularization4.7 x 10^17 FLOPEnhanceNet1.3 x 10^17 FLOPU-Net5.1 x 10^16 FLOPDiffDock-PP3.5 x 10^16 FLOPLSTM LM1.7 x 10^16 FLOPLCNP LabelMe3.3 x 10^15 FLOPPooling CNN (NORB)1.5 x 10^15 FLOPPooling CNN (Caltech 101)1.2 x 10^15 FLOPLSTM3.2 x 10^13 FLOPBiLSTM for Speech2.4 x 10^13 FLOPPredictive Coding NN1.9 x 10^13 FLOPFLUX.2November 2025FLUX.1 Kontext [max]May 2025FLUX.1 Kontext [dev]May 2025FLUX.1 Kontext [pro]May 2025RNAformerDecember 2024ConfRankNovember 2024FLUX1.1 [pro] UltraNovember 2024FLUX1.1 [pro] RawNovember 2024SO3LROctober 2024SPOTSeptember 2024ProteinGeneratorSeptember 2024JaegerSeptember 2024Automated design of multi-target ligands by generative deep learningSeptember 2024KinoMLSeptember 2024VespaGSeptember 2024BetterBodiesSeptember 2024Flux.1 [dev]August 2024Flux.1 [pro]August 2024DeepL LLMJuly 2024MetaMath 7B (Mistral finetune)May 2024MetaMath 7B (LLaMa finetune)May 2024MetaMath 70BMay 2024HiFi - NNDecember 2023Bird Vocalization Classifier (Perch)December 2023SILC-SOctober 2023OpenFlamingoAugust 2023AlphaLinkApril 2023
GPU clusters (11)
"Jupiter, Jülich"23.5K H100eJUWELS-Booster1.2K H100eFZJ JURECA242 H100eMax-Planck-Gesellschaft Raven242 H100eKarlsruher Institut für Technologie HoreKa233 H100ehessian.AI fortytwo199 H100eAleph Alpha alpha ONE161 H100eTU Dresden AlphaCentauri85.76 H100eUniversitaet Frankfurt Goethe-NHR80.49 H100eHigh-Performance Computing Center Stuttgart Hawk60.54 H100eContinental DGX Supercomputer28.3 H100e
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 countries