This document is relevant for: Inf1, Inf2, Trn1, Trn2, Trn3
AWS Neuron SDK 2.33.0 release notes
Date of release: October 19, 2026
This page provides detailed component release notes for the Neuron SDK 2.33.0. For an overview of the release content, see What’s New in AWS Neuron.
Package and Library Updates
The libraries and packages updated in this Neuron release.
Component Release Notes
Select a card below to review detailed release notes for each component of the Neuron SDK version 2.33.0. These component release notes contain details on specific new and improved features, as well as breaking changes, bug fixes, and known issues for that component area of the Neuron SDK. Some components may not be updated in this release.
PyTorch framework integration for AWS Trainium.
TorchTitan extension for large-model pre-training on AWS Trainium with TorchNeuron.
JAX framework integration for AWS Trainium and Inferentia.
Neuron’s Python-based programming interface for developing and optimizing Neuron kernels.
Reference kernels and utilities for Neuron kernel development with NKI.
The Neuron Graph Compiler that transforms ML graphs and NKI kernels into Neuron Executable File Format (NEFF) files.
The Neuron kernel driver and C++ libraries for AWS Inferentia and Trainium instances.
Tools that support end-to-end development for AWS Neuron, including Neuron Explorer.
Low-overhead cluster health and observability for Trainium.
Open-source AI agents and skills for NKI kernel authoring, debugging, profiling, and HuggingFace model porting in agentic IDEs.
High-performance LLM inference serving on Neuron with vLLM.
Diffuser-type multimodal generation (image and video) on Neuron with vLLM Omni.
AWS-specific machine images for building and deploying Neuron-based ML solutions.
AWS-specific container definitions for building and deploying Neuron-based ML solutions.
Previous releases
This document is relevant for: Inf1, Inf2, Trn1, Trn2, Trn3
