Nvidia’s AI agent game is here with new models and coordination schemes

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… deadLlama and were trained on the company’s technologies and datasets. The company also announced new AI orchestration schemes to guide AI agents. These latest releases put Nvidia, the company best known for the hardware powering the generative AI revolution, at the forefront of agentic AI development.

The Nemotron is available in three sizes: Nano, Super, and Ultra. It also comes in two flavors: the Llama Nemotron for language tasks and the Cosmos Nemotron vision model for physical AI projects. The Llama Nemotron Nano has 4B parameters, Super 49B parameters, and Ultra 253B parameters.

They all work best for agented tasks including “following instructions, 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 were optimized to fit different Nvidia computing resources. Nano is for low-latency, cost-effective applications on PCs and peripherals, while Super is for high resolution and throughput on a single GPU and Ultra for the highest resolution at data center scale.

“AI agents are the digital workforce that will work for us and work with us, so the Nemotron model family is dedicated to agent AI,” said Leparidian.

Nemotron models are available as hosted APIs on Hugging Face and the Nvidia website. Nvidia said companies can access the models through its AI Enterprise software platform.

Nvidia is no stranger to foundation models. Last year, it was released quietly Version of Nemotron, Llama-3.1-Nemotron-70B-Instructwhich outperformed similar models of OpenAI and Anthropic. It is too NVLM 1.0 has been detected,A family of multimodal language models.

More support for agents

Artificial intelligence agents It is becoming a big trend in 2024 as organizations start exploring how to deploy proxy systems in their workflow. Many think so The momentum will continue this year.

Companies like Sales force, Service now, Os and Microsoft They have all called agents the next wave of artificial general intelligence in enterprises. AWS added Multi-agent format to Bedrock, while Salesforce released a Agent Force 2.0bringing more agents to its clients.

However, agent workflows still need other infrastructure to operate efficiently. One such infrastructure revolves around orchestration, or managing multiple agents crossing different systems.

Organization charts

Nvidia has also entered the emerging field of AI orchestration with its blueprints that guide customers through specific tasks.

The company has partnered with several coordination companies, including langshen, LlamaIndex, CrewAI, daily and Weights and biasesto create charts on Nvidia AI Enterprise. Each orchestration framework has developed its own schema using Nvidia. For example, CrewAI created a code documentation schema to ensure code repositories are easy to navigate. LangChain has added Nvidia NIM microservices to its structured report generation scheme to help agents return internet searches in different formats.

“Getting multiple agents to work together seamlessly or coordinate is key to deploying agentic AI,” Libaridian said. “These leading AI orchestration companies are integrating all of the Nvidia, NIM, Nemo and Blueprints proxy building blocks with their open source proxy orchestration platforms.”

Nvidia’s new PDF to podcast conversion scheme aims to compete with Google NotebookLM By converting information from PDF files to audio. Another new scheme will help build agents to search and summarize videos.

Blueprints aims to help developers deploy AI agents quickly, Leparidian said. To that end, Nvidia has unveiled Nvidia Launchables, a platform that lets developers test blueprints, create prototypes, and launch them with one click.

The format can be one of Bigger stories for 2025 As companies grapple with multi-agent production.

Daily insights into business use cases with VB Daily

If you want to impress your boss, VB Daily has you covered. We give you the inside scoop on what companies are doing with generative AI, from organizational transformations to hands-on deployments, so you can share insights to maximize ROI.

Read our privacy policy

Thanks for subscribing. Check more VB newsletters here.

An error occurred.