Nemotron 3-Nano-30B-A3B
Nemotron 3-Nano-30B-A3B is an open-weights language model from NVIDIA, released December 15, 2025. It has 30B parameters (mixture-of-experts, 3B active).
We present Nemotron 3 Nano 30B-A3B, a Mixture-of-Experts hybrid MambaTransformer language model. Nemotron 3 Nano was pretrained on 25 trillion text tokens, including more than 3 trillion new unique tokens over Nemotron 2, followed by supervised fine tuning and large-scale RL on diverse environments.
Check it against your own PC →
Can you run Nemotron 3-Nano-30B-A3B locally?
The smallest common GPU it runs on is 6 GB of VRAM (with 32 GB of RAM): Q6_K, GPU + experts in RAM, roughly 18 tokens per second. Estimates assume 32 GB of DDR5 RAM, a 16K context and llama.cpp-style GGUF quantization. Enter your exact hardware in the live app.
| GPU memory | Verdict | Best quant | VRAM | RAM | tok/s |
|---|---|---|---|---|---|
| 6 GB VRAM | Runs · GPU + CPU experts | Q6_K (≈ lossless) | 5.2 GB | 22.6 GB | ~18 |
| 8 GB VRAM | Runs · GPU + CPU experts | Q6_K (≈ lossless) | 7.2 GB | 20.6 GB | ~20 |
| 12 GB VRAM | Runs · GPU + CPU experts | Q6_K (≈ lossless) | 11.2 GB | 16.6 GB | ~24 |
| 16 GB VRAM | Runs · GPU + CPU experts | Q6_K (≈ lossless) | 15.2 GB | 12.6 GB | ~30 |
| 24 GB VRAM | Runs on GPU | Q4_K_M (very good) | 21.5 GB | – | ~230 |
| 32 GB VRAM | Runs on GPU | Q6_K (≈ lossless) | 27.8 GB | – | ~300 |
Benchmark results
Epoch AI has not published results for Nemotron 3-Nano-30B-A3B yet. New models are usually evaluated within days to weeks.
Links
More from NVIDIA
NVIDIA Nemotron Labs Dory1.0 4B BF16 · NVIDIA Nemotron Labs Dory1.0 4B MXFP8 · Nemotron 3.5 Lightning · Nemotron 3 Ultra · Nemotron 3 Super · Nemotron-Cascade 14B · NVIDIA Isaac GR00T N1.6 · NVIDIA-Nemotron-Nano-9B-v2