In the press8 September 2026

Pathway’s brain-inspired architecture development on Amazon SageMaker HyperPod

AWS Machine Learning BlogOriginally published on aws.amazon.com

AWS Machine Learning Blog on how Pathway develops and scales Baby Dragon Hatchling (BDH), a brain-inspired post-transformer architecture that reasons in latent space instead of emitting chain-of-thought tokens, on Amazon SageMaker HyperPod - and how BDH-CQ set a new cost-efficiency mark on ARC-AGI-1 at 29.2% pass@2 for $0.0007 per task.

This article was published by AWS Machine Learning Blog. Pathway is featured in it - the full piece lives on their site.