NVIDIA has released Nemotron 3.5 Lightning, a 30B-parameter open mixture-of-experts model with 3B active parameters, alongside NeMo Switchyard, an open source routing library for multi-model AI agent workflows. The model targets the execution layer of long-running agents, handling tool calls, result validation, and subagent delegation more efficiently than frontier models. Its hybrid Mamba-2, MoE, and Attention architecture supports a 1M-token context window. Licensed under OpenMDW-1.1 with open weights and data, Lightning is deployable on a single modern GPU and via cloud providers. Early adopters span cybersecurity, legal, coding, finance, and healthcare, expecting lower costs and faster throughput.
This update represents a notable development in the Ai sector. Organizations and founders tracking this space should evaluate potential strategic and technical implications on their operations.