GPT-Neo Models by EleutherAI
0 models
Last refreshed 2026-10-10. Next refresh: weekly.
About
GPT-Neo is EleutherAI’s family of autoregressive Transformer language models based on its replication of the GPT-3 architecture. EleutherAI released 1.3B and 2.7B models trained on the Pile on March 21, 2021, and a 125M model is also available; the 2.7B model was trained on 420 billion tokens. The models are general-purpose text generators that continue a prompt. The original GPT-Neo codebase, built on Mesh TensorFlow, is now archived, and EleutherAI moved its development to GPT-NeoX.
Decision facts
- Best fit
- coding
- Capability starting point
- Not tracked
- Lowest tracked input
- Not tracked
- Closest related family
- GPT-NeoX
