OpenAI Five Rerun

Training compute

1.3×10²² FLOP

Parameters

159M

Published

Dec 13, 2019

OpenAI Five Rerun is an AI model developed by OpenAI (United States), first published in December 2019. It works in the games domain, on tasks such as dota 2.

Training it took an estimated 1.3×10²² FLOP of compute (estimation method: third-party estimation). The model has 159,000,000 parameters. It was trained on roughly 53.1B datapoints. Training ran on 512 NVIDIA P100. The compute alone is estimated at $321K in 2023 dollars.

Access: Unreleased. Its weights are not openly released. The reference paper has 2,137 citations. Epoch AI rates the confidence of this record as confident.

Full record

Organization: OpenAI
Country of organization: United States
Domain: Games
Task: Dota 2
Training compute: 1.3×10²² FLOP
Compute estimation method: Third-party estimation
Parameters: 159,000,000
Dataset size: 53.1B
Training hardware: NVIDIA P100
Chips used: 512
Training power draw: 262.1 kW
Training cost (2023 USD): $321K
Model accessibility: Unreleased
Open weights: No
Citations: 2,137
Epoch confidence: Confident