Nemotron 3 Ultra NVIDIA's Beast Model
In this video, we look at the latest Nemotron model from NVIDIA, Nemotron 3 Ultra. This is a 550 billion parameter model with 55 billion parameters active. #NVIDIAAI #Nemotron3Ultra Blog: https://nvda.ws/3PTkjlQ HF: https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-NVFP4 Paper: https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Ultra-Technical-Report.pdf Twitter: https://x.com/Sam_Witteveen 🕵️ Interested in building LLM Agents? Fill out the form below Building LLM Agents Form: https://drp.li/dIMes 👨💻Github: https://github.com/samwit/llm-tutorials ⏱️Time Stamps: 00: 00 Intro 00:13 Nemotron 3 Ultra 00:52 NVIDIA's Nemotron 3 Family - Nano, Super, Ultra 01:28 Model Specs - 550B, MoE, Agentic Focus 04:05 Multi-Teacher Distillation 05:29 Post-Training for Agent Harnesses 06:17 RL Environments 08:07 Benchmarks 08:56 Pinchbench 10:14 Live Demo - NVIDIA Cloud API Walkthrough 10:34 Reasoning Modes 12:50 Tool Calling
