LaunchAI ModelsAI Agents13 days agoNVIDIA launches Nemotron 3 Ultra, a 550B open model for agentsNVIDIA Developer Blog550B MoE with 55B active parameters, hybrid Mamba-Attention architecture, 1M context. Achieves 5x throughput and 30% lower cost for long-running agent tasks compared to peers.NemotronNvidiaHugging FaceSee 14 more sourcesnvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP413 days agonvidianvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF1613 days agonvidiaAmerica's leading open-source model, Nemotron 3 Ultra, is now available on Perplexity for all Pro...12 days agoAravind SrinivasNemotron 3 Ultra is now available for Pro and Max subscribers on Perplexity and Computer. It's...12 days agoPerplexity AIRT @alvarobartt: NVIDIA Nemotron 3 Ultra is here to supercharge your agents! 550B MoE, 55B active...13 days agoHugging FaceRT @HuggingPapers: NVIDIA just released Nemotron 3 Ultra on Hugging Face 550B total params, 55B...13 days agoHugging FaceNemotron 3 Ultra and Latent MoE Scaling19 hours agoSebastian RaschkaIntroducing NVIDIA Nemotron 3 Ultra13 days agoNVIDIANemotron 3 Ultra now available on AI Gateway13 days agoJerilyn ZhengNemotron 3 Ultra: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning2 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 MiNvidia’s best model is now live13 days agoFrederic LardinoisAsk the Experts: Nemotron 3 Ultra | Nemotron Labs12 days agoNVIDIA DeveloperNemotron 3 Ultra Tutorial: Build an Autonomous Research Agent with NemoHermes and OpenCode12 days agoNVIDIA DeveloperIntroducing Nemotron 3 Ultra13 days agoNVIDIA DeveloperNVIDIA Developer Blog·Chris Alexiuk13 days ago
