NVIDIA Unveils Nemotron Model Families to Advance AI Agents and Enterprise Productivity

NVIDIA Introduces Nemotron Model Families for Advanced AI Agents

NVIDIA has unveiled its latest innovation in artificial intelligence: the Nemotron model families, including Llama Nemotron large language models (LLMs) and Cosmos Nemotron vision language models (VLMs). Announced at CES 2025, these new models are designed to propel the development of agentic AI, ushering in a new era of enterprise productivity and problem-solving capabilities 1

1

2

2

.

Llama Nemotron: Powering Next-Generation AI Agents

Built on Meta's Llama foundation models, the Llama Nemotron family is optimized for creating and deploying AI agents across various applications. These open-source LLMs excel in tasks such as instruction following, chat functionality, function calling, coding, and mathematics 1

1

3

3

. NVIDIA claims that the models are trained using "latest techniques and high-quality datasets," although specific architectural details were not disclosed 1

1

.

Cosmos Nemotron: Enhancing Visual AI Capabilities

Complementing the Llama Nemotron models, NVIDIA introduced Cosmos Nemotron VLMs. These models enable AI agents to analyze and respond to images and videos, making them suitable for deployment in autonomous machines, healthcare settings, retail environments, and media applications 1

1

3

3

.

Flexible Deployment Options

Both Nemotron families will be available in three parameter sizes:

1. Nano: Cost-effective and optimized for low-latency, real-time applications on edge devices and PCs.
2. Super: High-accuracy model offering exceptional throughput on a single GPU.
3. Ultra: Highest-accuracy model designed for data center-scale applications 2

   2

   4

   4

   .

Enterprise Integration and Customization

NVIDIA is offering the Nemotron models as downloadable resources and as NVIDIA NIM microservices, allowing for easy deployment across various computing platforms 2

2

. Enterprises can customize these models for specific use cases using NVIDIA NeMo microservices, which simplify data curation, accelerate model customization, and apply guardrails 3

3

.

Industry Support and Applications

Leading AI platform providers, including SAP and ServiceNow, have expressed support for the Nemotron models. SAP plans to incorporate them into its Joule platform to enhance enterprise user productivity, while ServiceNow aims to utilize the models for AI agent services across various industries 2

2

4

4

.

Availability and Access

The Llama Nemotron and Cosmos Nemotron models will soon be available through build.nvidia.com, Hugging Face, and the NVIDIA Developer Program. While the models are open-source, they are currently limited to academic and research usage 1

1

2

2

. Enterprise-grade deployments will be supported via the NVIDIA AI Enterprise platform on accelerated cloud and data center infrastructure 2

2

.

As AI continues to evolve, NVIDIA's Nemotron model families represent a significant step towards more sophisticated and capable AI agents, potentially revolutionizing how enterprises approach complex problem-solving and task automation across various industries.