RunLocalAI
other
30B parameters
Commercial OK

Nemotron 3 Nano (30B-A3B)

NVIDIA's hybrid Mamba-2 + Transformer MoE for on-device agents. 30B total / 3B active. 1M-token context window with reasoning ON/OFF modes and 4× faster inference than the previous Nemotron Nano.

License: NVIDIA Open Model License·Released Jan 15, 2026·Context: 1,000,000 tokens

Overview

NVIDIA's hybrid Mamba-2 + Transformer MoE for on-device agents. 30B total / 3B active. 1M-token context window with reasoning ON/OFF modes and 4× faster inference than the previous Nemotron Nano.

Strengths

  • 1M-token context
  • Reasoning toggle
  • Hybrid Mamba architecture

Weaknesses

  • Newer architecture — runner support varies

Quantization variants

Each quantization trades model quality for file size and VRAM. Q4_K_M is the most popular starting point.

QuantizationFile sizeVRAM required
Q4_K_M18.0 GB22 GB
Q8_032.0 GB36 GB

Get the model

Ollama

One-line install

ollama run nemotron3:nanoRead our Ollama review →

HuggingFace

Original weights

huggingface.co/nvidia/Nemotron-3-Nano

Source repository — direct quantization required.

Hardware that runs this

Cards with enough VRAM for at least one quantization of Nemotron 3 Nano (30B-A3B).

NVIDIA GB200 NVL72
13824GB · nvidia
AMD Instinct MI355X
288GB · amd
AMD Instinct MI325X
256GB · amd
AMD Instinct MI300X
192GB · amd
NVIDIA B200
192GB · nvidia
NVIDIA H100 NVL
188GB · nvidia
NVIDIA H200
141GB · nvidia
Intel Gaudi 3
128GB · intel
Compare alternatives

Models worth comparing

Same parameter band, plus what's one tier above and below — so you can decide what actually fits your hardware.

Frequently asked

What's the minimum VRAM to run Nemotron 3 Nano (30B-A3B)?

22GB of VRAM is enough to run Nemotron 3 Nano (30B-A3B) at the Q4_K_M quantization (file size 18.0 GB). Higher-quality quantizations need more.

Can I use Nemotron 3 Nano (30B-A3B) commercially?

Yes — Nemotron 3 Nano (30B-A3B) ships under the NVIDIA Open Model License, which permits commercial use. Always read the license text before deployment.

What's the context length of Nemotron 3 Nano (30B-A3B)?

Nemotron 3 Nano (30B-A3B) supports a context window of 1,000,000 tokens (about 1000K).

How do I install Nemotron 3 Nano (30B-A3B) with Ollama?

Run `ollama pull nemotron3:nano` to download, then `ollama run nemotron3:nano` to start a chat session. The default quantization is Q4_K_M.

Source: huggingface.co/nvidia/Nemotron-3-Nano

Reviewed by RunLocalAI Editorial. See our editorial policy for how we research and verify model claims.