AI models
Jais
Training compute
4.9×10²² FLOP
Parameters
13B
Published
Aug 29, 2023
Jais is an AI model developed by Cerebras Systems, Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) and Inception G42 (United States and United Arab Emirates), first published in August 2023. It works in the language domain, on tasks such as language modeling, chat, translation and 2 more.
Training it took an estimated 4.9×10²² FLOP of compute (estimation method: operation counting,hardware). The model has 13,000,000,000 parameters. It was trained on roughly 395B datapoints. Training ran on 16 Cerebras CS-2 for about 600 hours.
Access: Open weights (unrestricted). Its weights are openly available. The reference paper has 74 citations. Epoch AI rates the confidence of this record as confident.
Full record
- Organization
- Cerebras Systems, Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), Inception G42
- Country of organization
- United States, United Arab Emirates
- Domain
- Language
- Task
- Language modeling, Chat, Translation, Language modeling/generation, Question answering
- Training compute
- 4.9×10²² FLOP
- Compute estimation method
- Operation counting, Hardware
- Parameters
- 13,000,000,000
- Dataset size
- 395B
- Training hardware
- Cerebras CS-2
- Chips used
- 16
- Training time
- 600 h
- Model accessibility
- Open weights (unrestricted)
- Open weights
- Yes
- Citations
- 74
- 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