Back to AIBriefs
LaunchAI ModelsAI Agents

NVIDIA releases open-source Nemotron 3 Ultra model

NVIDIA releases open-source Nemotron 3 Ultra modelNVIDIA Developer Blog

Nemotron 3 Ultra is a 550B-parameter MoE model with 55B active parameters, featuring a hybrid Mamba-2 MoE Transformer and 1M context window. It achieves 5x higher throughput than comparable open models and reduces agentic task costs by up to 30%. Available on Perplexity Pro/Max and Hugging Face.

nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP414 days agonvidia
nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF1615 days agonvidia
America's leading open-source model, Nemotron 3 Ultra, is now available on Perplexity for all Pro...13 days agoAravind Srinivas
RT @NVIDIAAI: Today we're shipping Nemotron 3 Ultra. A 550B MoE frontier-intelligence open model...15 days agoHugging Face
RT @TheAhmadOsman: Models I am excited to run locally these next few months - Nemotron 3 Ultra -...21 days agoHugging Face
RT @alvarobartt: NVIDIA Nemotron 3 Ultra is here to supercharge your agents! 550B MoE, 55B active...15 days agoHugging Face
Nemotron 3 Ultra and Latent MoE Scaling2 days ago
Introducing NVIDIA Nemotron 3 Ultra15 days agoNVIDIA
Nemotron 3 Ultra now available on AI Gateway15 days agoJerilyn Zheng
Nemotron 3 Ultra: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning3 days agoNVIDIA (Allan), : (Allan), Aaron Blakeman (Allan), Aaron Thomas (Allan), Aastha Jhunjhunwala (Allan), Abhibha Gupta (Allan), Abhinav Khattar (Allan), Adam Rajfer (Allan), Adi Renduchintala (Allan), Adil Asif (Allan), Aditya Vavre (Allan), Adriana Flores Mi
Nvidia’s best model is now live15 days agoFrederic Lardinois
Nemotron 3 Ultra Tutorial: Build an Autonomous Research Agent with NemoHermes and OpenCode14 days agoNVIDIA Developer
Nemotron Office Hours: The Nemotron 3 Model Family | Nemotron Labs11 hours agoNVIDIA Developer
Ask the Experts: Nemotron 3 Ultra | Nemotron Labs13 days agoNVIDIA Developer
·
Jun 4, 1:02 PM