Nvidia's new Nemotron 3 family combines Mamba and Transformer architectures to handle long context windows without burning through resources.
Skip to content
Dec 17, 2025
Nvidia
Nvidia's new Nemotron 3 family combines Mamba and Transformer architectures to handle long context windows without burning through resources.
The new model generation targets agent-based AI, systems that autonomously handle complex tasks over extended periods. The lineup includes three models: Nano, Super, and Ultra. Nano is available now, while Super and Ultra are scheduled for the first half of 2026.
NVIDIA has released Nemotron 3 Ultra, a 550B total (55B active) open Mixture-of-Experts hybrid Mamba-Transformer for long-running…
As the industry shifts from chatbots to multi-agent workflows, Nvidia's Nemotron 3 offers a blueprint for efficient, long-context…
A new, open, 120-billion-parameter hybrid mixture-of-experts model optimized for NVIDIA Blackwell addresses the costs of long…
NVIDIA Nemotron 3 Super is now available on Together AI Dedicated Inference, delivering efficient multi-agent reasoning, a…
Nemotron 3 Nano, NVIDIA’s newest reasoning model, is now available on Together AI, the AI Native Cloud
Best-in-class open omni-modal reasoning model delivers the highest efficiency and accuracy to power agentic workflows such as…
