NVIDIA releases open-source AI model Nemotron 3.5 Lightning, 4x faster

NVIDIA launched the 30B open-source AI model Nemotron 3.5 Lightning on August 11. It uses a MoE design and is built for specific tasks in large multi-agent systems, aiming to boost the development and performance of agent applications.

Compared to similar models, Nemotron 3.5 Lightning delivers up to 4x faster output, speeding up overall agent task completion by 30%. It runs on local AI systems like RTX PCs and DGX Spark, and can also scale to enterprise edge devices, data centers, and cloud setups for flexible deployment.

The model is now available for free on Hugging Face, ModelScope, OpenRouter, and NVIDIA's site. This move expands the Nemotron 3 series, offering a more efficient open-source option for AI agents.