AI models
HAMSTER VLM
Training compute
2.4×10²¹ FLOP
Parameters
13.5B
Published
Feb 8, 2025
HAMSTER VLM is an AI model developed by NVIDIA, University of Washington and University of Southern California (United States), first published in February 2025. It works in the robotics domain, on tasks such as robotic manipulation.
Training it took an estimated 2.4×10²¹ FLOP of compute (estimation method: hardware). The model has 13,493,916,736 parameters. Training ran on 8 NVIDIA A100 SXM4 80 GB for about 30 hours.
Access: Open weights (non-commercial). Its weights are openly available. It is built on top of VILA1.5-13B. The reference paper has 104 citations. Epoch AI rates the confidence of this record as confident.
Full record
- Organization
- NVIDIA, University of Washington, University of Southern California
- Country of organization
- United States
- Domain
- Robotics
- Task
- Robotic manipulation
- Training compute
- 2.4×10²¹ FLOP
- Compute estimation method
- Hardware
- Parameters
- 13,493,916,736
- Training hardware
- NVIDIA A100 SXM4 80 GB
- Chips used
- 8
- Training time
- 30 h
- Training power draw
- 6.3 kW
- Numerical format
- BF16
- Model accessibility
- Open weights (non-commercial)
- Open weights
- Yes
- Base model
- VILA1.5-13B
- Citations
- 104
- 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