NVIDIA has announced Nemotron 3, a family of open models, data and librariesdesigned to drive transparent, efficient and specialized development of AI with agents in all sectors. Arriving in Nano, Super and Ultra sizes, the models introduce an innovative hybrid latent mixture of experts (MoE) architecture that helps developers build and deploy reliable multi-agent systems at scale.
As organizations shift from single-model chatbots to multi-agent collaborative AI systems, developers face increasing challenges such as communication overhead, context drift, and high inference costs, NVIDIA says. Additionally, developers need transparency to trust the models that will automate their complex workflows. Nemotron 3 directly addresses these challenges, delivering the performance and openness customers need to build specialized AI with agents.
NVIDIA Nemotron supports NVIDIA’s broader initiatives in IA sovereign. Organizations from Europe to South Korea are adopting open, transparent and efficient models that allow them to build AI systems aligned with their own data, regulations and values.
Pioneering companies, including Accenture, Cadence, CrowdStrike, Cursor, Deloitte, EY, Oracle Cloud Infrastructure, Palantir, Perplexity, ServiceNow, Siemens, Synopsys and Zoom, are integrating models from the Nemotron 3 family to power AI workflows in industries such as manufacturing, cybersecurity, software development, media, communications and more.
Nemotron 3, “reinventing multi-agent AI”
Nemotron 3’s open models allow large organizations and startups to develop and Interact faster on AI agents and accelerate innovation from prototype to enterprise deployment. The idea is to create AI teams that facilitate collaboration between humans and artificial intelligence.
The Nemotron 3 family of MoE models includes three sizes:
- Nemotron 3 Nanoa small 30 billion parameter model that activates up to 3 billion parameters at a time for specific and highly efficient tasks.
- Nemotron 3 Supera high-precision reasoning model with approximately 100 billion parameters and up to 10 billion assets per token, for multi-agent applications.
- Nemotron 3 Ultraa large reasoning engine with around 500 billion parameters and up to 50 billion assets per token, for complex AI applications.
Nemotron 3 Nano is now available from Hugging Face and through inference service providers such as Baseten, DeepInfra, Fireworks, FriendliAI, OpenRouter and Together AI. Nemotron is offered on enterprise AI and data infrastructure platforms such as Couchbase, DataRobot, H2O.ai, JFrog, Lambda, and UiPath. For customers on public clouds, Nemotron 3 Nano will be available on AWS through Amazon Bedrock (serverless), as well as supporting Google Cloud, CoreWeave, Crusoe, Microsoft Foundry, Nebius, Nscale, and Yotta coming soon.
Nemotron 3 Nano is available as an NVIDIA NIM microservice for secure, scalable deployment anywhere on NVIDIA-accelerated infrastructure for maximum privacy and control. Nemotron 3 Super and Ultra are expected to be available in the first half of 2026.
More information | NVIDIA Nemotron 3
