NVIDIA Nemotron 3.5 Lightning available on SageMaker JumpStart for agents

AWS ML Blog · rss · 2026-08-18

NVIDIA Nemotron 3.5 Lightning is now available on Amazon SageMaker JumpStart. Featuring a hybrid MoE architecture (30B total, 3B active) and a 1M token context window, the model is optimized for high-volume agentic workflows. NVIDIA reports up to 4x higher throughput and 30% faster task completion, making it suitable for running specialized steps like classification and extraction on single GPUs.

Original post →

More from coding & agent

coding & agent channel →