NVIDIA releases Nemotron 3 Ultra, 550B-param open MoE model

Vercel Blog

Nemotron 3 Ultra is a 550B-parameter (55B active) hybrid Mamba-2 MoE transformer with 1M token context. It achieves up to 350 tokens/s and 30% lower cost on agentic tasks. The open-weight model is available on Hugging Face and Vercel AI Gateway.

NvidiaNemotron

See 11 more sources

nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF1618 days agonvidia

RT @alvarobartt: NVIDIA Nemotron 3 Ultra is here to supercharge your agents! 550B MoE, 55B active...18 days agoHugging Face

RT @NVIDIAAI: Today we're shipping Nemotron 3 Ultra. A 550B MoE frontier-intelligence open model...18 days agoHugging Face

RT @HuggingPapers: NVIDIA just released Nemotron 3 Ultra on Hugging Face 550B total params, 55B...18 days agoHugging Face

RT @mervenoyann: NVIDIA Nemotron Ultra is here 😍 > 55B/550B a hybrid MoE  🦖 with 1M context...18 days agoHugging Face

1/ NVIDIA shipped Nemotron 3 Ultra today, a fully open 550B model with 55B active params, with the...18 days agoKimmonismus

NVIDIA just announced the release of Nemotron 3 Ultra in Jensen Huang's Computex keynote: at 550B...22 days agoArtificial Analysis

NVIDIA has just released Nemotron 3 Ultra, the new most intelligent US open weights model, with...18 days agoArtificial Analysis

Introducing NVIDIA Nemotron 3 Ultra: An Open 550B Model for Long-Running Agents18 days agoNVIDIA Developer

Nemotron 3 Ultra Tutorial: Build an Autonomous Research Agent with NemoHermes and OpenCode17 days agoNVIDIA Developer

Nemotron 3 Ultra and the Open Model Landscape | Nemotron Labs20 hours agoNVIDIA Developer

Vercel Blog·Jerilyn Zheng

Jun 4, 7:00 AM