NVIDIA Releases Nemotron 3 Ultra for Enhanced Agentic AI
NVIDIA has released Nemotron 3 Ultra, a 550B-parameter Mixture-of-Experts model, optimized for orchestrating complex, long-running agent workflows. It features architectural innovations such as hybrid Mamba-Transformer layers for efficient long-context handling, NVFP4 quantization for up to 5x higher throughput, LatentMoE for expert routing, and multi-token prediction for improved generative speed. The model is fully open with weights, data, and recipes, and trained with Multi-Teacher On-Policy Distillation.
Nemotron 3 Ultra provides a powerful and open platform for developers to build and deploy highly efficient and adaptable AI agents, leveraging advanced architectural designs and training methodologies.
Learn one new AI thing every day.
Daily Deck sends you seven plain-English cards like this every morning. Free.
Start free