Const Launches Teutonic-II: 110B Permissionless Training Run on Bittensor Subnet 3
The new Subnet 3 run scales Teutonic’s competition-based training mechanism from a 10B model to a 110B sparse Mixture-of-Experts model.
Teutonic
Teutonic Subnet Begins Training 80B AI Model on Bittensor, Marking Largest Decentralized Training Run Yet
Teutonic Subnet 3 has begun training an 80B AI model on Bittensor, marking the largest decentralized AI training effort yet and a major milestone for TAO infrastructure.