NVIDIA releases Nemotron 3 Ultra, a 550B MoE open model

Hugging Face Models

Nemotron 3 Ultra is a 550B total (55B active) hybrid Mamba-Transformer MoE model, released as open weights. It is designed for long-running agents and is the largest Nemotron 3 model, announced at Computex by Jensen Huang.

NvidiaNemotronOpen source

See 13 more sources

Creating the NVIDIA Nemotron 3 Ultra NVFP4 Checkpoint with NVIDIA Model Optimizer13 hours agoMichelle Horton

RT @alvarobartt: NVIDIA Nemotron 3 Ultra is here to supercharge your agents! 550B MoE, 55B active...22 days agoHugging Face

RT @NVIDIAAI: Today we're shipping Nemotron 3 Ultra. A 550B MoE frontier-intelligence open model...22 days agoHugging Face

RT @HuggingPapers: NVIDIA just released Nemotron 3 Ultra on Hugging Face 550B total params, 55B...22 days agoHugging Face

Nemotron 3 Ultra and Latent MoE Scaling9 days ago

NVIDIA just announced the release of Nemotron 3 Ultra in Jensen Huang's Computex keynote: at 550B...26 days agoArtificial Analysis

NVIDIA has just released Nemotron 3 Ultra, the new most intelligent US open weights model, with...22 days agoArtificial Analysis

1/ NVIDIA shipped Nemotron 3 Ultra today, a fully open 550B model with 55B active params, with the...22 days agoKimmonismus

Nemotron 3 Ultra now available on AI Gateway22 days agoJerilyn Zheng

Nemotron 3 Ultra Tutorial: Build an Autonomous Research Agent with NemoHermes and OpenCode21 days agoNVIDIA Developer

Ask the Experts: Nemotron 3 Ultra | Nemotron Labs21 days agoNVIDIA Developer

Introducing NVIDIA Nemotron 3 Ultra: An Open 550B Model for Long-Running Agents22 days agoNVIDIA Developer

Introducing Nemotron 3 Ultra22 days agoNVIDIA Developer

Hugging Face Models·nvidia

Jun 4, 4:59 PM