ALOHA Unleashed

Training compute

3.6×10²¹ FLOP

Parameters

217M

Published

Sep 8, 2024

ALOHA Unleashed is an AI model developed by Google DeepMind (United States), first published in September 2024. It works in the robotics domain, on tasks such as robotic manipulation.

Training it took an estimated 3.6×10²¹ FLOP of compute (estimation method: hardware). The model has 217,000,000 parameters. Training ran on 64 Google TPU v5e for about 265 hours.

Access: Unreleased. Its weights are not openly released. Epoch AI rates the confidence of this record as confident.

Full record

Organization: Google DeepMind
Country of organization: United States
Domain: Robotics
Task: Robotic manipulation
Training compute: 3.6×10²¹ FLOP
Compute estimation method: Hardware
Parameters: 217,000,000
Training hardware: Google TPU v5e
Chips used: 64
Training time: 265 h
Training power draw: 28.4 kW
Model accessibility: Unreleased
Open weights: No
Epoch confidence: Confident