What is Nemotron 3 Nano 30B A3B good for?

Agent systems, chatbots and RAG in English and code, on local and edge hardware. Open weights under the NVIDIA Nemotron Open Model License, on a single consumer GPU.

When is Nemotron 3 Nano 30B A3B not a good fit?

Image input: it is text only. The NVIDIA Nemotron Open Model License has its own terms rather than Apache 2.0 or MIT, and no maximum output is published.

What is the cheapest way to run Nemotron 3 Nano 30B A3B?

Hosted, unless you push serious volume. Geodd charges $0.05 in / $0.20 out per 1M tokens. The cheapest rental that fits is 2x RTX 3060 at $72 a month, which costs the same as about 960M tokens a month on that API.

Can I self-host Nemotron 3 Nano 30B A3B?

Yes. The weights are released under the NVIDIA Nemotron Open Model License. At 4-bit it needs about 20 GB of GPU memory, which starts at roughly $72 a month on the cheapest rental that fits.

Open weights · NVIDIA Nemotron Open Model License

Nemotron 3 Nano 30B A3B

The smallest of NVIDIA's Nemotron 3 line, a 30B hybrid Mamba-2, MoE and attention model activating 3.5B parameters per token under the NVIDIA Nemotron Open Model License, sized for local and edge use.

Jump to: Specs Pricing Self-hosting Similar models FAQ More from Nvidia

Visit website Compare vs other LLMs →

Cheapest via Geodd

Input: $0.05 / 1M tokens
Output: $0.20 / 1M tokens

4 providers

Key Specifications

Context window: 262K tokens
Knowledge cutoff: Jun 2025
Released: Dec 2025
Parameters: 31.6B, 3.6B active
Inputs: Text

Capabilities Show details: Reasoning Thinks before it answers, always on or as a switchable mode. Function calling Connect to external tools, APIs, and systems. Structured output Return responses in structured formats like JSON.

Hosted API pricing

Input tokens M Output tokens M

Provider	Input / 1M tokens	Output / 1M tokens	Cost at 10M in + 2M out
Geodd	$0.05	$0.20	$0.90	View
Novita	$0.05	$0.20	$0.90	View
Lyceum	$0.06	$0.24	$1.08	View
Nvidia Creator	Open weights, no hosted price			View

Heads up: Base-tier, on-demand rates per 1M tokens; cached, batch and long-context tiers excluded. A provider may serve a shorter context or a quantized build than the creator's release. Verify before provisioning. More on how we price.

Estimated cost to self-host

Nemotron 3 Nano 30B A3B needs about 20 GB of GPU memory at 4-bit with 32K context. The cheapest rental that fits is 2× RTX 3060 at about $72 a month.

That costs the same as roughly 960M tokens a month on Geodd's API. Self-hosting is more expensive below that volume.

Precision	Cheapest, 32K context	Cheapest, full 262K context
4-bitINT4 / FP4	2× RTX 3060 20 GB VRAM · $72/mo	2× RTX 3060 21 GB VRAM · $72/mo
8-bitFP8 / INT8	4× RTX 3060 34 GB VRAM · $144/mo	4× RTX 3060 35 GB VRAM · $144/mo
16-bitFP16 / BF16	8× RTX 3060 65 GB VRAM · $288/mo	8× RTX 3060 67 GB VRAM · $288/mo

Estimates based on median on-demand rates for Nvidia GPUs. Memory is weights plus KV cache for one request, using FP8 KV cache where supported and FP16/BF16 otherwise. Break-even assumes a 5:1 input-to-output ratio. No guarantee of runtime support, usable performance, or that a matching quantized build exists. How we estimate costs.

Similarly priced models

The models nearest Nemotron 3 Nano 30B A3B by blended rate, each at its own cheapest provider.

Model	Blended / 1M	Input / 1M	Output / 1M	Context	Cutoff	vs Nemotron 3 Nano 30B A3B
GPT-OSS-20B OpenAI	$0.0483	$0.03	$0.14	131K	Jun 2024	−36%
Command R7B Cohere	$0.0583	$0.04	$0.15	128K		−22%
Gemma 3 12B Instruct Google Cloud	$0.0583	$0.05	$0.10	128K	Aug 2024	−22%
Llama 3.2 1B Meta	$0.0583	$0.03	$0.20	128K	Dec 2023	−22%
GPT-OSS-120B OpenAI	$0.0633	$0.04	$0.18	131K	Jun 2024	−16%
Nemotron 3 Nano 30B A3B This model Nvidia	$0.075	$0.05	$0.20	262K	Jun 2025
Qwen3-Coder-30B-A3B Alibaba Cloud	$0.0917	$0.06	$0.25	262K		+22%
Llama 3.2 3B Meta	$0.0983	$0.05	$0.34	128K	Dec 2023	+31%
Qwen3-30B-A3B FP8 Alibaba Cloud	$0.0983	$0.05	$0.34	33K		+31%
Ministral 3 3B Mistral	$0.10	$0.10	$0.10	256K		+33%
DeepSeek V4 Flash DeepSeek	$0.1083	$0.08	$0.25	1M		+44%

Prices are USD per 1M tokens at each model's cheapest listed provider. Blended is the cost of 10M input plus 2M output tokens, spread over the 12M.

Thunder Compute

Our sponsor

NVIDIA H100 from $1.38/hr, A100 80GB from $0.78/hr, RTX A6000 from $0.27/hr

No commitments, billed per minute, SOC 2 and GDPR compliant

Hot swappable hardware, templates for ComfyUI and Ollama

One-click connection to VS Code or Cursor, simple CLI to connect without SSH

Try now

Frequently Asked Questions

What is Nemotron 3 Nano 30B A3B good for?: Agent systems, chatbots and RAG in English and code, on local and edge hardware. Open weights under the NVIDIA Nemotron Open Model License, on a single consumer GPU.
When is Nemotron 3 Nano 30B A3B not a good fit?: Image input: it is text only. The NVIDIA Nemotron Open Model License has its own terms rather than Apache 2.0 or MIT, and no maximum output is published.
What is the cheapest way to run Nemotron 3 Nano 30B A3B?: Hosted, unless you push serious volume. Geodd charges $0.05 in / $0.20 out per 1M tokens. The cheapest rental that fits is 2x RTX 3060 at $72 a month, which costs the same as about 960M tokens a month on that API.
Can I self-host Nemotron 3 Nano 30B A3B?: Yes. The weights are released under the NVIDIA Nemotron Open Model License. At 4-bit it needs about 20 GB of GPU memory, which starts at roughly $72 a month on the cheapest rental that fits.

More from Nvidia

Model	Context	Cutoff	Input / 1M	Output / 1M
Nemotron 3 Super 120B A12B Open weights	1M	Jun 2025	$0.09	$0.50
Nemotron 3 Ultra 550B A55B Open weights	1M	Sep 2025	$0.66	$2.64