NVIDIA Unveils Nemotron 3.5 Lightning and NeMo Switchyard for Efficient Agentic AI

NVIDIA is expanding its Nemotron 3 model family with Nemotron 3.5 Lightning, which the company describes as the highest-efficiency model in its class for long-running agentic AI workloads, according to a post on the NVIDIA Blog.

The release arrives as AI shifts from chatbots to autonomous agents, and as open models increasingly serve demand for full control over where AI runs and how it is deployed and evolves.

Nemotron 3.5 Lightning is positioned as the efficiency-focused addition to the Nemotron 3 line, aimed at workloads where agents must run for extended periods without excessive compute cost.

The announcement also introduces NeMo Switchyard, NVIDIA's tooling for building and deploying agentic AI systems, pairing the new model with the software stack needed to put it to work.

The launch is tied to NVIDIA's push for local and on-premises AI, with the release covering both RTX and DGX platforms so users can run open models in environments they control.

Why it matters: efficiency is becoming the deciding factor for agentic AI, where long-running loops multiply inference cost, and a model that maintains quality at lower overhead changes the economics of deploying agents at scale.

The move also strengthens NVIDIA's position in the open-model ecosystem, where local AI communities are building, customizing, and running increasingly capable agents on their own hardware.

What to watch next: benchmark results against other open agentic models, adoption within NVIDIA's NeMo ecosystem, and how the broader Nemotron 3 family evolves in the coming months.