Country
Switzerland
Tracked models
63
Frontier models
2
Largest training run
6.7 x 10^24 FLOP
63 tracked AI models have at least one contributing organization based in Switzerland. 2 of them are frontier models — among the largest training runs of their moment. The largest is Apertus 70B, at 6.7 x 10^24 FLOP (September 2025). The most recent tracked release is from September 2025.
It hosts 4 tracked GPU clusters.
Models (63)
Apertus 70B6.7 x 10^24 FLOPMamayLM4.4 x 10^23 FLOPSILC-S* (86M)10^22 FLOPYaRN (Llama 2 70B)10^21 FLOPDiamond8 x 10^20 FLOPDiffSBDD (CrossDocked)2.7 x 10^20 FLOPYaRN (Llama 2 13B)1.9 x 10^20 FLOPRoBERTa (PFAM)1.2 x 10^19 FLOPEquiDock1.1 x 10^19 FLOPLinear Transformer (large)3.9 x 10^18 FLOPRMSNorm (Transformer-base)2.8 x 10^18 FLOPMoLeR2.1 x 10^18 FLOPLinear Transformer (small)1.7 x 10^18 FLOPDelta RNN (+ full context)1.1 x 10^18 FLOPAWD-FWM (WT2)7.1 x 10^17 FLOPSelf-Attention and Convolutional Layers6.8 x 10^17 FLOPDNN EM segmentation4.8 x 10^17 FLOPAWD-LSTM-DRILL + dynamic evaluation† (WT2)4.1 x 10^17 FLOPMitosis1.4 x 10^17 FLOPVariational RHN + WT (PTB)1.3 x 10^17 FLOPCNN Committee (MNIST)5.2 x 10^16 FLOPCNN Committee (NIST)2.6 x 10^16 FLOPHigh Performance CNN (NORB)2.6 x 10^16 FLOPMCDNN (MNIST)1.6 x 10^16 FLOPVD-RHN3.6 x 10^15 FLOPGCRN-M1, dropout3 x 10^15 FLOPCNN committee (traffic sign)9.9 x 10^14 FLOP6-layer MLP (MNIST)1.3 x 10^14 FLOPBiLSTM for Speech2.4 x 10^13 FLOPApertus 8BSeptember 2025Boltz-2June 2025GenesisOctober 2024BindCraftOctober 2024YOLOv8 (reCAPTCHA fine-tuned)September 2024RENNAISSANCEAugust 2024Rethinking Molecular Design: Integrating Latent Variable and Auto-Regressive Models for Goal Directed GenerationAugust 2024ColPaliJuly 2024DiffPALMJune 2024MPNNsolJune 2024Meditron-70BNovember 2023DiT-XL/2 + CADSOctober 2023SILC-SOctober 2023SignalP 6.0January 2022PersiaNovember 2021CloobOctober 2021Fairseq + UID: varianceMay 2021AWD-FWM (PTB)November 2020StarGAN v2December 2019Alleviated TOI 10 (WT103)September 2019Alleviated TOI 10 (WT2)September 2019Alleviated TOI 10 (PTB)September 2019PG-SWGANJune 2019AWD-LSTM-DRILL + dynamic evaluation† (PTB)May 2019TSNSeptember 2016Highway NetworkNovember 2015Clockwork RNN (CW-RNN)February 2014Semi-Supervised Embedding for DLJuly 2008BLSTM for handwriting (2)December 2007BLSTM for handwriting (1)September 2007CTC-Trained LSTMJune 2006Peephole LSTMJuly 2000LSTM with forget gatesJanuary 1999Samuel Neural Checkers IINovember 1967
GPU clusters (4)
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