AI models
Octo-Base
Training compute
5.8×10²⁰ FLOP
Parameters
93M
Published
May 20, 2024
Octo-Base is an AI model developed by University of California (UC) Berkeley, Stanford University, Carnegie Mellon University (CMU) and DeepMind (United States and United Kingdom), first published in May 2024. It works in the robotics domain, on tasks such as robotic manipulation.
Training it took an estimated 5.8×10²⁰ FLOP of compute (estimation method: hardware). The model has 93,000,000 parameters. Training ran on 128 Google TPU v4 for about 14 hours.
Access: Open weights (unrestricted). Its weights are openly available. The reference paper has 1,244 citations. Epoch AI rates the confidence of this record as confident.
Full record
- Organization
- University of California (UC) Berkeley, Stanford University, Carnegie Mellon University (CMU), DeepMind
- Country of organization
- United States, United Kingdom
- Domain
- Robotics
- Task
- Robotic manipulation
- Training compute
- 5.8×10²⁰ FLOP
- Compute estimation method
- Hardware
- Parameters
- 93,000,000
- Training hardware
- Google TPU v4
- Chips used
- 128
- Training time
- 14 h
- Training power draw
- 86.0 kW
- Numerical format
- BF16
- Model accessibility
- Open weights (unrestricted)
- Open weights
- Yes
- Citations
- 1,244
- Epoch confidence
- Confident
More from University of California (UC) Berkeley,Stanford University,Carnegie Mellon University (CMU),DeepMind
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