DeepHorus is an innovative tech media and intelligence platform focused on AI, blockchain, startups, and emerging technologies.
Published on

NVIDIA launches 30B open-source AI model Nemotron 3.5 Lightning

Authors
  • avatar
    Name
    DeepHorus
    Twitter

On August 11, local time, NVIDIA announced the launch of Nemotron 3.5 Lightning, an open-source AI model with 30 billion parameters, specifically designed for specific tasks in large-scale multi-agent systems. The release further enriches NVIDIA's Nemotron 3 model series, providing developers with more efficient agent solutions.

With the rapid development of AI agent applications, developers have an increasing demand for efficient, lightweight models. Nemotron 3.5 Lightning adopts a MoE (Mixture of Experts) architecture, and compared with similar models, its output speed can be increased by up to 4 times, thereby accelerating the overall completion speed of agent tasks by 30%. This performance breakthrough enables the model to respond more quickly in complex reasoning and interaction tasks.

In terms of deployment, Nemotron 3.5 Lightning supports direct operation on local AI systems such as NVIDIA RTX PC, DGX Spark, DGX Station, and Jetson. At the same time, it can seamlessly scale to edge AI devices, RTX PRO workstations, data centers, and cloud environments, meeting the multi-level needs from individual developers to enterprise applications.

NVIDIA has opened model downloads on platforms such as Hugging Face, ModelScope, and OpenRouter, and developers can also use it online through the NVIDIA Build platform. This open-source strategy is expected to attract more developers to participate in the construction of the agent ecosystem and promote innovation in AI applications.

Industry analysts believe that the launch of this model lowers the threshold for multi-agent applications and accelerates the evolution of AI from single models to collaborative agents. In particular, the substantial increase in output speed provides stronger support for real-time decision-making and automated processes in enterprises.

In the future, NVIDIA may continue to improve the Nemotron series, optimizing model performance and deployment compatibility. With the growth of the open-source ecosystem, agent applications based on Nemotron 3.5 Lightning are expected to be first implemented in fields such as customer service, programming, and simulation, helping the AI industry develop to a higher level.