Nemotron
18 models · NVIDIA
NVIDIA's open Nemotron family — Llama-derived and original models (Nano ~4–12B, Super ~49–120B, Ultra up to 550B MoE) heavily optimized via pruning, distillation and reasoning post-training, plus vision (VL) and omni variants. Designed to showcase NVIDIA hardware and ship with open training data recipes.
Models in this family
- Llama 3.1 Nemotron Instruct 70BNVIDIA70B parametersOpen weights
- Llama 3.1 Nemotron Nano 4B v1.1 (Reasoning)NVIDIA4B parametersOpen weights
- Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)NVIDIA253B parametersOpen weights
- Llama 3.3 Nemotron Super 49B v1 (Non-reasoning)NVIDIA49B parametersOpen weights
- Llama 3.3 Nemotron Super 49B v1 (Reasoning)NVIDIA49B parametersOpen weights
- Llama Nemotron Super 49B v1.5 (Non-reasoning)NVIDIA49B parametersOpen weights
- Llama Nemotron Super 49B v1.5 (Reasoning)NVIDIA49B parametersOpen weights
- Nemotron 3 Nano Omni 30B A3B ReasoningNVIDIA30B total / 3B active (MoE)Open weights
- Nemotron 3 Ultra 550B A55B (Reasoning)NVIDIA550B total / 55B active (MoE)Open weights
- Nemotron Cascade 2 30B A3BNVIDIA30B total / 3B active (MoE)Open weights
- NVIDIA Nemotron 3 Nano 30B A3B (Non-reasoning)NVIDIA30B total / 3B active (MoE)Open weights
- NVIDIA Nemotron 3 Nano 30B A3B (Reasoning)NVIDIA30B total / 3B active (MoE)Open weights
- NVIDIA Nemotron 3 Nano 4BNVIDIA4B parametersOpen weights
- NVIDIA Nemotron 3 Super 120B A12B (Reasoning)NVIDIA120B total / 12B active (MoE)Open weights
- NVIDIA Nemotron Nano 12B v2 VL (Non-reasoning)NVIDIA12B parametersOpen weights
- NVIDIA Nemotron Nano 12B v2 VL (Reasoning)NVIDIA12B parametersOpen weights
- NVIDIA Nemotron Nano 9B V2 (Non-reasoning)NVIDIA9B parametersOpen weights
- NVIDIA Nemotron Nano 9B V2 (Reasoning)NVIDIA9B parametersOpen weights