AI models
SPHINX (Llama 2 13B)
Training compute
3×10²² FLOP
Parameters
19.9B
Published
Nov 13, 2023
SPHINX (Llama 2 13B) is an AI model developed by Shanghai AI Lab, Chinese University of Hong Kong (CUHK) and ShanghaiTech University (China and Hong Kong), first published in November 2023. It works in the vision, language and multimodal domain, on tasks such as visual question answering and image captioning.
Training it took an estimated 3×10²² FLOP of compute (estimation method: hardware). The model has 19,900,000,000 parameters. Training ran on 32 NVIDIA A100 SXM4 40 GB for about 290 hours. The compute alone is estimated at $239K in 2023 dollars.
Access: Open weights (restricted use). Its weights are openly available. It is built on top of Llama 2-13B. Epoch AI rates the confidence of this record as likely.
Full record
- Organization
- Shanghai AI Lab, Chinese University of Hong Kong (CUHK), ShanghaiTech University
- Country of organization
- China, Hong Kong
- Domain
- Vision, Language, Multimodal
- Task
- Visual question answering, Image captioning
- Training compute
- 3×10²² FLOP
- Compute estimation method
- Hardware
- Parameters
- 19,900,000,000
- Training hardware
- NVIDIA A100 SXM4 40 GB
- Chips used
- 32
- Training time
- 290 h
- Chip-hours
- 9.3K
- Training power draw
- 25.4 kW
- Training cost (2023 USD)
- $239K
- Numerical format
- BF16
- Model accessibility
- Open weights (restricted use)
- Open weights
- Yes
- Base model
- Llama 2-13B
- 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