Organization

AT&T

Tracked models

18

Frontier models

3

Largest training run

2.8 x 10^12 FLOP

18 tracked AI models list AT&T as a developer. 3 of them are frontier models — among the largest training runs of their moment. The largest is LeNet-5, at 2.8 x 10^12 FLOP (November 1998). The most recent tracked release is from September 2009.

Models (18)

LeNet-52.8 x 10^12 FLOP Zip CNN1.5 x 10^12 FLOP Handwritten digit recognition network1.8 x 10^11 FLOP BellKor 2007September 2009 MatrixFac for RecommendersAugust 2009 BigChaos OptiBlendAugust 2009 BellKor 2008August 2009 BellKor 2009August 2009 BigChaos 2008November 2008 Enhanced Neighborhood-Based FilteringOctober 2007 Tagging via Viterbi DecodingJune 2002 Perceptron for Large Margin ClassificationDecember 1999 Learning to Order ThingsMay 1999 Social and content-based classificationJuly 1998 AdaBoost.M2 Digit RecognitionJuly 1996 Support Vector MachinesSeptember 1995 Random Decision ForestsAugust 1995 Learning-curve predictionNovember 1993

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.