NVIDIA Nemotron 3.5 Lightning now available in Amazon SageMaker JumpStart
NVIDIA Nemotron 3.5 Lightning brings 4x agentic throughput to SageMaker JumpStart on a single GPU
NVIDIA Nemotron 3.5 Lightning, a 30B-total/3B-active MoE model distilled from Nemotron 3 Ultra, is now deployable via Amazon SageMaker JumpStart without manual infrastructure setup. It delivers up to 4x throughput and 30% faster task completion on agentic workloads, with a 1M-token context window, all on a single GPU. The launch advances the 'system-of-models' architecture pattern where high-volume, repetitive agent steps are routed to efficient specialized models rather than costly frontier models.