AlphaTensor

Training compute

7.1×10²⁰ FLOP

Published

Oct 5, 2022

AlphaTensor is an AI model developed by DeepMind (United Kingdom), first published in October 2022. It works in the other, games and mathematics domain.

Training it took an estimated 7.1×10²⁰ FLOP of compute (estimation method: hardware). Training ran on 64 Google TPU v3 for about 168 hours.

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

Full record

Organization: DeepMind
Country of organization: United Kingdom
Domain: Other, Games, Mathematics
Training compute: 7.1×10²⁰ FLOP
Compute estimation method: Hardware
Training hardware: Google TPU v3
Chips used: 64
Training time: 168 h
Training power draw: 57.6 kW
Model accessibility: Unreleased
Open weights: No
Epoch confidence: Confident