AI models
RetNet
Training compute
4×10²¹ FLOP
Parameters
6.7B
Published
Jul 17, 2023
RetNet is an AI model developed by Microsoft Research and Tsinghua University (United States and China), first published in July 2023. It works in the language domain, on tasks such as language modeling.
Training it took an estimated 4×10²¹ FLOP of compute (estimation method: operation counting). The model has 6,700,000,000 parameters. It was trained on roughly 100B datapoints.
Access: Unreleased. Its weights are not openly released. The reference paper has 646 citations. Epoch AI rates the confidence of this record as confident.
Full record
- Organization
- Microsoft Research, Tsinghua University
- Country of organization
- United States, China
- Domain
- Language
- Task
- Language modeling
- Training compute
- 4×10²¹ FLOP
- Compute estimation method
- Operation counting
- Parameters
- 6,700,000,000
- Dataset size
- 100B
- Model accessibility
- Unreleased
- Open weights
- No
- Citations
- 646
- Epoch confidence
- Confident
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.
← All ai models