Nemotron 3 Nano 4B: A Compact Hybrid Model for Efficient Local AI

Training Recipe for Nemotron 3 Nano 4B Compressing 9B → 4B with Nemotron Elastic How the Router Decides What to Prune Two-Stage Distillation for Accuracy Recovery Supervised Fine-Tuning Multi-environment Reinforcement Learning Boosting Efficiency with Quantization Try It Now! We are excited to introduce Nemotron 3 Nano 4B, the newest and most compact member of the Nemotron 3 family. Leveraging hybrid Mamba-Transformer architecture, this model is designed for efficiency and accuracy in a targeted