AI models
ALLaM adapted 70B
Training compute
1.1×10²⁴ FLOP
Parameters
70B
Published
May 21, 2024
ALLaM adapted 70B is an AI model developed by Saudi Data and Artificial Intelligence Authority (Saudi Arabia), first published in May 2024. It works in the language domain, on tasks such as language modeling/generation, translation and question answering.
Training it took an estimated 1.1×10²⁴ FLOP of compute. The model has 70,000,000,000 parameters. It was trained on roughly 600B datapoints. Training ran on NVIDIA A100.
Access: Unreleased. Its weights are not openly released. It is built on top of Llama 2-70B. The reference paper has 69 citations. Epoch AI rates the confidence of this record as confident.
Full record
- Organization
- Saudi Data and Artificial Intelligence Authority
- Country of organization
- Saudi Arabia
- Domain
- Language
- Task
- Language modeling/generation, Translation, Question answering
- Training compute
- 1.1×10²⁴ FLOP
- Parameters
- 70,000,000,000
- Dataset size
- 600B
- Training hardware
- NVIDIA A100
- Numerical format
- BF16
- Model accessibility
- Unreleased
- Open weights
- No
- Base model
- Llama 2-70B
- Citations
- 69
- Epoch confidence
- Confident
More from Saudi Data and Artificial Intelligence Authority
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