Nvidia’s New Open Model Is for Specific Use Cases

Amid heightened interest in Chinese open models, AI hardware vendor Nvidia introduced its latest open source model: Nemotron 3.5 Lightning.Released on August 11, Nemotron 3.5 Lightning is a mixture-of-experts model that is designed for specialized tasks within larger multi-agent systems. Nemotron 3.5 Lightning can perform specific tasks such as code review, tool use, security alert monitoring and answering billing questions, Nvidia said. The vendor also released the model weights, code and recipes and training under the Linux Foundation’s OpenMDW 1.1 license, making it a truly open source model.Nemotron 3.5 Lightning is part of Nvidia’s Nemotron 3 family of open models that was originally released last December.The model has 30 billion parameters, which is smaller than Nemotron 3 Ultra (550 billion) and larger than Nemotron 3 Nano (8 billion). The model is meant to help enterprises create agentic applications on their local devices.Related:Meta Reverses Course with Open-Weight Muse GlimmerAlong with Nemotron 3.5 Lightning, Nvidia released NeMo Switchyard, an open source library for routing inside popular agentic applications. Enterprises can use Switchyard to route a prompt or agentic request to the most suitable model across their own mix of open, proprietary or Nvidia models, the AI hardware vendor said.Both Nemotron 3.5 Lightning and NeMo Switchyard were released less than a month after Nvidia CEO Jensen Huang revealed that the AI hardware-software giant had joined a large group of tech vendors in defense of open source as the Trump administration weighed restrictions on open-weight models from China. In an X post, Huang said open models enable sovereignty and strengthen safety and cybersecurity.Nemotron 3.5 Lightning is Nvidia’s first open model since debate about advancements in Chinese open models ramped up last month, after Chinese AI vendors Alibaba and Moonshot AI released cheap, powerful open-weight models. It comes a day after social media giant Meta released its latest open model.“It’s a testament to just how impactful the Chinese models have been and will be,” said Bradley Shimmin, an analyst at Futurum Group, adding that many consider Alibaba Qwen 3.8 Max to be “the benchmark for local capable open source agentic development.”Supporting Nvidia’s InfrastructureHowever, for Nvidia, providing open source models is not only a strategy to compete against Chinese vendors but also a way to help sell its chips and hardware, according to Arun Chandrasekaran, an analyst at Gartner.“Models help Nvidia sell its infrastructure better,” Chandrasekaran said. “When you sell a model, you have to run the model somewhere. Running the models requires networking, inference software and data training.Related:Microsoft Framework to Cut AI Agent Training Costs“It is in Nvidia’s own interest that there is a vibrant model ecosystem that’s out there, and the model ecosystem is not very concentrated,” Chandrasekaran continued.He added that, for the most part, enterprises are only using Nvidia’s models for specific tasks or when an enterprise wants to deploy models on premises or where there are significant sovereignty considerations.With Nemotron 3.5 Lightning, Nvidia said enterprise users can run the model on local AI systems, including Nvidia’s RTX PCs, DGX Spark and Jetson.Training Data and Model SizeThat Nvidia also released the training data for the model is also commendable, Shimmin said.“There are very few vendors, with the exception of IBM, that have been promoting and touting this ability to say, ‘if you indeed care about exposure to risk in adopting AI, you should care about the training data used to build these models,’” he said. “It all starts with the data, and it honestly ends with the data too.”The size of Nemotron 3.5 Lightning illustrates how vendors like Nvidia are beginning to acknowledge the need to “right-size” models now, Shimmin added. He said that many vendors realize that the current way of using APIs to access models on a token per cost basis is too expensive, and so being able to work with the model locally helps enterprises trying to justify the cost of generative AI.Related:Oracle Brings Google Gemini Models to Enterprise Customers“Tokenomics over time have already proven to be a little bit hazardous for companies in terms of both not understanding their token spend and having no ability to actually gain observability or transparency in how that spend is going,” he said.About the AuthorNews Writer, AI BusinessEsther Shittu has covered AI technologies and industry trends since 2021. As co-host of the Targeting AI podcast, she talks with experts, thought leaders and practitioners exploring critical AI developments. Before AI Business, she wrote for SearchEnterpriseAI, the New York Daily News, Bklyner and the Brooklyn Daily Eagle. When she's not diving deep into the world of AI, she spends her time on passion projects and raising her three daughters.