Nemotron-3 Nano Omni 30B-A3B
MoENVIDIA Open License
Hybrid Mamba-2 + MoE + Attention architecture. Native unified four-modality (text/image/video/audio) understanding; best-in-class on MMLongBench-Doc / OCRBenchV2 / VoiceBench. Mamba layers give ~4× compute efficiency. 256K context (up to 1M). BF16 on HuggingFace, Unsloth GGUF and free OpenRouter tier available.
Provider
NVIDIA
Parameters
3.5B active / 30B total (Mamba-2 + MoE + Attention)
Context
262.144K
Released
2026-04-29
VRAM Requirements by Quantization
| Method | Disk Size | VRAM Required | Fits GPUs |
|---|---|---|---|
| Q8_0 | 31 GB | 33 GB | 4 GPUs |
| Q4_K_M | 17.5 GB | 19 GB | 9 GPUs |
| Q4_0 | 16.5 GB | 18 GB | 9 GPUs |
| Q2_K | 10.5 GB | 12 GB | 16 GPUs |
Install with Ollama
Run in terminal:
ollama pull nemotron3-nano-omniMinimum 12GB VRAM required. Install Ollama from ollama.com
Benchmark Scores
mmlu80.5%
humaneval78%
Scores are approximate and may vary by quantization level.
Compatible GPUs (16)
AMD RX 9070 XT (16GB)AMD RX 7900 GRE (16GB)AMD RX 7900 XTX (24GB)AMD Ryzen AI Max+ 395 (unified memory) (64GB)Apple M4 Pro (24GB) (24GB)Apple M3 Max (36GB) (36GB)Apple M4 Max (48GB) (48GB)Apple M4 Ultra (64GB) (64GB)NVIDIA RTX 4070 Ti SUPER (16GB)NVIDIA RTX 4080 SUPER (16GB)NVIDIA RTX 5070 Ti (16GB)NVIDIA RTX 4060 Ti 16GB (16GB)NVIDIA RTX 5080 (16GB)NVIDIA RTX 4090 (24GB)NVIDIA RTX 3090 (24GB)NVIDIA RTX 5090 (32GB)
HuggingFace
nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning