AI models
CodeT5-base
Training compute
1.6×10²¹ FLOP
Parameters
220M
Published
Nov 1, 2021
CodeT5-base is an AI model developed by Salesforce and Nanyang Technological University (United States and Singapore), first published in November 2021. It works in the language domain, on tasks such as code generation.
Training it took an estimated 1.6×10²¹ FLOP of compute (estimation method: hardware). The model has 220,000,000 parameters. Training ran on NVIDIA A100 for about 288 hours. The compute alone is estimated at $3K in 2023 dollars.
Access: Open weights (unrestricted). Its weights are openly available. The reference paper has 1,964 citations. Epoch AI rates the confidence of this record as likely.
Full record
- Organization
- Salesforce, Nanyang Technological University
- Country of organization
- United States, Singapore
- Domain
- Language
- Task
- Code generation
- Training compute
- 1.6×10²¹ FLOP
- Compute estimation method
- Hardware
- Parameters
- 220,000,000
- Training hardware
- NVIDIA A100
- Training time
- 288 h
- Training cost (2023 USD)
- $3K
- Numerical format
- FP16
- Model accessibility
- Open weights (unrestricted)
- Open weights
- Yes
- Citations
- 1,964
- Epoch confidence
- Likely
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