Nemotron 3 Nano 30B MoE: Open Generative AI on SageMaker JumpStart
The NVIDIA Nemotron 3 Nano 30B MoE model is now generally available in the Amazon SageMaker JumpStart catalog, offering a powerful solution for accelerating generative AI applications on AWS. This small language hybrid Mixture of Experts (MoE) model, featuring 30 billion total parameters with 3 billion active parameters, is designed for exceptional compute efficiency and accuracy, making it ideal for developers tackling highly-skilled agentic tasks at scale. A significant benefit is its fully open nature, providing open-weights, datasets, and recipes, which allows developers to customize, optimize, and deploy the model on their own infrastructure, thereby meeting specific privacy and security requirements.
Nemotron 3 Nano boasts a unique architecture combining MoE with a hybrid Transformer-Mamba design, and supports a token budget for optimal accuracy with minimal reasoning token generation. It exhibits leading accuracy across critical benchmarks in coding, scientific reasoning, math, and instruction following, outperforming other open language models under 30B in areas like SWE Bench Verified, GPQA Diamond, and AIME 2025. With a vast context window of up to 1 million tokens and text-based input/output, it offers high usability for complex tasks.
Targeting developers and businesses, Nemotron 3 Nano on SageMaker JumpStart simplifies model deployment, removing management complexities and enabling rapid innovation. Users with an Amazon SageMaker Studio domain can easily discover, deploy, and test the model through the SageMaker Studio interface, or programmatically via AWS CLI, SageMaker SDK, or Boto3. This integration democratizes access to advanced AI, allowing customers to build production-ready generative AI applications and leverage Nemotron’s capabilities efficiently. The model’s availability underscores AWS and NVIDIA’s commitment to providing accessible, high-performance AI tools.
Amazon’s ai automation sagemaker platform now supports Nemotron 3 Nano 30B MoE, enabling developers to deploy advanced generative AI models with streamlined workflows.
Organizations seeking alternatives to ChatGPT automation SageMaker solutions can now leverage Nemotron 3 Nano’s advanced capabilities for enterprise AI deployments.
