AI models
CHAI-1
Training compute
7.8×10²¹ FLOP
Published
Oct 15, 2024
CHAI-1 is an AI model developed by Chai discovery (United States), first published in October 2024. It works in the biology domain, on tasks such as protein folding prediction and protein-ligand contact prediction.
Training it took an estimated 7.8×10²¹ FLOP of compute (estimation method: hardware). Training ran on 128 NVIDIA A100 for about 720 hours.
Access: Open weights (non-commercial). Its weights are openly available. Epoch AI rates the confidence of this record as confident.
Full record
- Organization
- Chai discovery
- Country of organization
- United States
- Domain
- Biology
- Task
- Protein folding prediction, Protein-ligand contact prediction
- Training compute
- 7.8×10²¹ FLOP
- Compute estimation method
- Hardware
- Training hardware
- NVIDIA A100
- Chips used
- 128
- Training time
- 720 h
- Training power draw
- 100.8 kW
- Numerical format
- BF16
- Model accessibility
- Open weights (non-commercial)
- Open weights
- Yes
- Citations
- 0
- 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