AI models
QT-Opt
Training compute
1.4×10¹⁹ FLOP
Parameters
1.2M
Published
Jun 27, 2018
QT-Opt is an AI model developed by Google Brain and University of California (UC) Berkeley (United States), first published in June 2018. It works in the robotics and vision domain, on tasks such as robotic manipulation.
Training it took an estimated 1.4×10¹⁹ FLOP of compute (estimation method: hardware). The model has 1,200,000 parameters. It was trained on roughly 11.6M datapoints. Training ran on NVIDIA P100 for about 104 hours. The compute alone is estimated at $1K in 2023 dollars.
Access: Unreleased. Its weights are not openly released. The reference paper has 1,692 citations. Epoch AI rates the confidence of this record as likely.
Full record
- Organization
- Google Brain, University of California (UC) Berkeley
- Country of organization
- United States
- Domain
- Robotics, Vision
- Task
- Robotic manipulation
- Training compute
- 1.4×10¹⁹ FLOP
- Compute estimation method
- Hardware
- Parameters
- 1,200,000
- Dataset size
- 11.6M
- Training hardware
- NVIDIA P100
- Training time
- 104 h
- Training cost (2023 USD)
- $1K
- Model accessibility
- Unreleased
- Open weights
- No
- Citations
- 1,692
- Epoch confidence
- Likely
More from Google Brain,University of California (UC) Berkeley
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