NVIDIA Releases Nemotron 3.5 Lightning, an Open-Weight Model for Multi-Agent Systems

NVIDIA has released a lightweight, open-weight AI model called Nemotron 3.5 Lightning. The company said the new model can run on a single graphics processing unit on a PC. Nemotron 3.5 Lightning launches weeks after NVIDIA and other tech companies urged the U.S. government to support open-weights.

What Capabilities Does Nemotron 3.5 Have?

NVIDIA’s new open-source AI model is designed to support specialized workloads within larger multi-agent systems. The chipmaker, which released three cheaper open-source models last year, said companies can download the 30-billion parameter AI model for free. They can modify and redistribute it without paying or requesting NVIDIA’s permission.

“The model delivers up to 4x faster output speed, leading to 30% faster agentic task completion compared with other models in its class. And because it’s open and customizable, Nemotron 3.5 Lightning can be easily post-trained with NVIDIA NeMo on an organization’s own domain data, tools and workflows to improve accuracy for specialized tasks,” NVIDIA stated in a blog post.

Nemotron 3.5 Lightning can support a range of NVIDIA systems, including DGX Spark, RTX PCs, and DGX Station. The chipmaker said the open-weight model can scale across cloud environments and data centers. Already, companies like CrowdStrike, Harvey, and CodeRabbit have tested and customized Nemotron 3.5 Lightning. NVIDIA reportedly distilled the new open-source model from its larger Nemotron models to boost its capabilities.

Debate on Open-Weight AI Models Continues

NVIDIA released the Nemotron single GPU model at a time when major tech companies continue to discuss open-weight AI. On August 10, Meta CEO Mark Zuckerberg supported open-weights in a long post as the company launched an open-source model called Muse Glimmer.

In the U.S., policymakers have raised concerns over open-source AI, questioning whether Chinese-based open-weights like DeepSeek and Kimi K3 should be subjected to sanctions. Last month, major American tech companies, including NVIDIA, signed the “Open Weights and American AI Leadership” letter to endorse open-weights.

“Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty,” Huang posted on X at the time.

The Opportunity in Open-Weights

Nvidia views the rise of open-source AI as an opportunity to drive chip demand. Nemotron 3.5 Lighting signals a shift toward AI models that are cheaper to deploy and easier to customize. As more companies experiment with open-weight AI, demand will likely expand beyond hyperscale data centers to PCs, workstations, and enterprise infrastructure.

With AI models becoming more accessible, AI workloads will likely increase. This will increase demand for the GPUs needed to run them. As competition intensifies, NVIDIA’s ability to combine hardware with capable AI models could make it more competitive.