NVIDIA Nemotron 3 Ultra on Amazon SageMaker JumpStart

Announces day-zero availability of NVIDIA’s Nemotron 3 Ultra model on SageMaker JumpStart featuring 550B total and 55B active parameters with a hybrid Transformer-Mamba architecture. The open model delivers 5x faster inference and up to 30% lower cost for agentic workloads through one-click deployment with million-token context length support for multi-step reasoning tasks.