AI models
LongVU
Parameters
7B
Published
Oct 22, 2024
LongVU is an AI model developed by Meta AI, King Abdullah University of Science and Technology (KAUST) and Korea University (United States, Saudi Arabia and South Korea), first published in October 2024. It works in the multimodal, video and language domain, on tasks such as video compression, video, language modeling/generation and video description.
Epoch AI has no training-compute estimate for this model. The model has 7,000,000,000 parameters. Training ran on 64 NVIDIA H100 SXM5 80GB.
Access: Open weights (unrestricted). Its weights are openly available. It is built on top of Qwen2-7B. The reference paper has 259 citations. Epoch AI rates the confidence of this record as confident.
Full record
- Organization
- Meta AI, King Abdullah University of Science and Technology (KAUST), Korea University
- Country of organization
- United States, Saudi Arabia, South Korea
- Domain
- Multimodal, Video, Language
- Task
- Video compression, Video, Language modeling/generation, Video description
- Parameters
- 7,000,000,000
- Training hardware
- NVIDIA H100 SXM5 80GB
- Chips used
- 64
- Training power draw
- 88.2 kW
- Model accessibility
- Open weights (unrestricted)
- Open weights
- Yes
- Base model
- Qwen2-7B
- Citations
- 259
- 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