Join our daily and weekly newsletters for the latest updates and exclusive content covering cutting-edge AI. Learn more
The industry’s push toward agentic AI continues, with Nvidia announcing several new services and models to facilitate the creation and deployment of AI agents.
Today, Nvidia launched Nemotron, a family of models based on Meta‘s Llama and trained in the company’s techniques and datasets. The company also announced new AI orchestration plans to guide AI agents. These latest releases put Nvidia, a company best known for the hardware that powers the generative AI revolution, at the forefront of agentic AI development.
Nemotron is available in three sizes: Nano, Super and Ultra. It also comes in two versions: the Llama Nemotron for language tasks and the Cosmos Nemotron vision model for physical AI projects. The Llama Nemotron Nano has 4B settings, Super 49B settings, and Ultra 253B settings.
All three work best for agentic tasks, including “instruction following, chatting, calling functions, coding, and math,” according to the company.
Rev. Lebaredian, vice president of Omniverse and simulation technology at Nvidia, said in a briefing with reporters that the three sizes are optimized for different Nvidia computing resources. Nano is intended for cost-effective, low-latency applications on PCs and edge devices, Super is intended for high precision and throughput on a single GPU, and Ultra is intended for maximum precision at data center scale.
“AI agents are the digital workforce that will work for us and work with us, and so the Nemotron family of models is aimed at agentic AI,” Lebaredian said.
Nemotron models are available as APIs hosted on Hugging Face and the Nvidia website. Nvidia said businesses can access the models through its AI Enterprise software platform.
Nvidia is no stranger to foundation models. Last year he quietly released a version of Nemotron, Lama-3.1-Nemotron-70B-Instructwhich outperformed similar models of OpenAI And Anthropic. It is also unveiled NVLM 1.0a family of multimodal language models.
More support for agents
AI Agents has become a big trend in 2024 as businesses began to explore how to deploy agentic systems into their workflow. Many believe that the momentum will continue This year.
Companies like Sales force, ServiceNow, AWS And Microsoft have all called agents the next wave of AI generation in business. AWS added multi-agent orchestration at Bedrock, while Salesforce released its Agentforce 2.0bringing more agents to its customers.
However, agent workflows still require other infrastructure to operate effectively. One of these infrastructures revolves around the orchestration or management of multiple agents crossing different systems.
Orchestration plans
Nvidia has also entered the emerging field of AI orchestration with its blueprints that guide agents through specific tasks.
The company has partnered with several orchestration companies, including LangChain, LamaIndex, CrewAI, Every day And Weights and biasesto create plans on Nvidia AI Enterprise. Each orchestration framework has developed its own model with Nvidia. For example, CrewAI created a code documentation template to ensure easy navigation of code repositories. LangChain added Nvidia NIM microservices to its structured reporting model to help agents return internet searches in different formats.
“Having multiple agents work together in a fluid or orchestrating manner is key to deploying agentic AI,” Lebaredian said. “These leading AI orchestration companies integrate all of the agentic building blocks from Nvidia, NIM, Nemo and Blueprints, into their open source agentic orchestration platforms.”
Nvidia’s new PDF-to-podcast model aims to compete Google NotebookLM by converting information from PDFs to audio. Another new model will help create agents that can search and summarize videos.
Lebaredian said Blueprints aims to help developers quickly deploy AI agents. To that end, Nvidia unveiled Nvidia Launchables, a platform that lets developers test, prototype, and execute plans with just one click.
Orchestration could be one of the biggest stories of 2025 as companies grapple with multi-agent production.
