AI models
LLaVA-NeXT-34B (LLaVA-1.6)
Training compute
2.6×10²⁰ FLOP
Parameters
34.8B
Published
Jan 30, 2024
LLaVA-NeXT-34B (LLaVA-1.6) is an AI model developed by University of Wisconsin Madison, ByteDance, Nanyang Technological University and University of California (UC) Berkeley (United States, China and Singapore), first published in January 2024. It works in the multimodal, language and vision domain, on tasks such as visual question answering, chat and question answering.
Training it took an estimated 2.6×10²⁰ FLOP of compute (estimation method: hardware). The model has 34,750,000,000 parameters. It was trained on roughly 89.3M datapoints. Training ran on 32 NVIDIA A100 for about 24 hours.
Access: Open weights (unrestricted). Its weights are openly available. Epoch AI rates the confidence of this record as likely.
Full record
- Organization
- University of Wisconsin Madison, ByteDance, Nanyang Technological University, University of California (UC) Berkeley
- Country of organization
- United States, China, Singapore
- Domain
- Multimodal, Language, Vision
- Task
- Visual question answering, Chat, Question answering
- Training compute
- 2.6×10²⁰ FLOP
- Compute estimation method
- Hardware
- Parameters
- 34,750,000,000
- Dataset size
- 89.3M
- Training hardware
- NVIDIA A100
- Chips used
- 32
- Training time
- 24 h
- Chip-hours
- 768
- Training power draw
- 25.3 kW
- Model accessibility
- Open weights (unrestricted)
- Open weights
- Yes
- Epoch confidence
- Likely
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