AI models
Ankh_base
Training compute
2.6×10²¹ FLOP
Parameters
740M
Published
Jan 16, 2023
Ankh_base is an AI model developed by Technical University of Munich and Columbia University (Germany and United States), first published in January 2023. It works in the biology domain, on tasks such as protein generation, proteins, protein or nucleotide language model (plm/nlm) and 4 more.
Training it took an estimated 2.6×10²¹ FLOP of compute (estimation method: third-party estimation). The model has 740,000,000 parameters. It was trained on roughly 14B datapoints. Training ran on Google TPU v4. The compute alone is estimated at $2K in 2023 dollars.
Access: Open weights (non-commercial). Its weights are openly available. The reference paper has 159 citations. Epoch AI rates the confidence of this record as confident.
Full record
- Organization
- Technical University of Munich, Columbia University
- Country of organization
- Germany, United States
- Domain
- Biology
- Task
- Protein generation, Proteins, Protein or nucleotide language model (pLM/nLM), Protein contact and distance prediction, Protein classification, Protein localization prediction, Protein fold classification
- Training compute
- 2.6×10²¹ FLOP
- Compute estimation method
- Third-party estimation
- Parameters
- 740,000,000
- Dataset size
- 14B
- Training hardware
- Google TPU v4
- Training cost (2023 USD)
- $2K
- Model accessibility
- Open weights (non-commercial)
- Open weights
- Yes
- Citations
- 159
- Epoch confidence
- Confident
More from Technical University of Munich,Columbia University
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