NVIDIA Advances AI with Open Models for Agents, Robotics, and Healthcare

NVIDIA is expanding its suite of open AI models, introducing advancements in agentic AI, physical AI, and healthcare applications. These new models – including Nemotron 3, Isaac GR00T N1.7, Alpamayo 1.5, Cosmos 3, and BioNeMo Proteina-Complexa – aim to empower developers and scientists to build more intelligent systems capable of reasoning and acting in both digital and real-world environments. The announcements were made at the GTC conference on March 16, 2026.

AI Agents Powered by Nemotron 3

The NVIDIA Nemotron family now includes several omni-understanding models designed to enhance AI agents. These models deliver improved natural language conversations, complex reasoning, and advanced visual capabilities.

• Nemotron 3 Ultra: Offers 5x throughput efficiency using the NVFP4 format on NVIDIA Blackwell chips, ideal for AI-native applications like coding assistants and complex workflow automation.
• Nemotron 3 Omni: Integrates audio, vision, and language understanding, enabling AI agents to extract insights from videos and documents with greater accuracy.
• Nemotron 3 VoiceChat: Supports real-time conversations with simultaneous listening and responding, combining speech recognition, language processing, and text-to-speech in a single system.
• Nemotron Safety Models: Detect unsafe content across text and images, enhancing the trustworthiness of multimodal systems.

Companies like Automation Anywhere, CodeRabbit, CrowdStrike, Cursor, Factory, Distyl, Genspark, Perplexity, and ServiceNow are already deploying NVIDIA Nemotron models. Edison Scientific’s Kosmos platform, powered by Nemotron, reportedly compresses months of research into a single day for over 50,000 researchers. NVIDIA

Advancements in Physical AI

NVIDIA is also pushing the boundaries of physical AI with new foundation models and simulation tools for robotics and autonomous vehicles.

• NVIDIA Cosmos 3: A world foundation model designed to unify synthetic world generation, physical AI reasoning, and action simulation.
• NVIDIA Isaac GR00T N1.7: An open reasoning vision language action (VLA) model specifically built for humanoids, now commercially viable for real-world deployment.
• NVIDIA Alpamayo 1.5: A reasoning VLA model that enhances autonomous vehicle reasoning with navigation guidance and flexible multi-camera support.

LG Electronics and NEURA are adopting GR00T N1.7 for scaling humanoid robot deployment. NVIDIA

Accelerating Healthcare and Life Sciences Research

NVIDIA’s BioNeMo platform is expanding as an open AI development platform for healthcare and life sciences, accelerating biomedical research, drug discovery, and medical imaging.

• Proteina-Complexa: A generative model for protein binder design that accelerates structure-based drug discovery.

NVIDIA has collaborated with EMBL’s European Bioinformatics Institute, Google DeepMind, and Seoul National University to expand the AlphaFold Protein Structure Database, adding approximately 30 million protein complex predictions. NVIDIA

The introduction of nvQSP, a GPU-accelerated simulation engine, aims to speed up pharmaceutical research by enabling faster and more comprehensive simulations of treatment scenarios. NVIDIA

Availability

Select NVIDIA open models, data, and frameworks are available on GitHub and Hugging Face, as well as through various cloud and AI infrastructure platforms and build.nvidia.com. Many models are also accessible as NVIDIA NIM™ microservices for secure, scalable deployment. NVIDIA