NVIDIA Nemotron 3 Nano 30B A3B is a small language MoE model with highest compute efficiency and accuracy for developers to build specialized agentic AI systems. The model is fully...
- Modalities
- text → text
- In / out per 1M
- Free / Free
- Context
- 256K tokens
- Added
- Dec 14, 2025
- Tokenizer
- Other
- Reasoning
- optional
Pricing
Per 1M tokens. The provider price and our flat 3% fee are separate columns — what you pay is their sum.
| Per 1M tokens | Provider | + 3% fee | You pay |
|---|---|---|---|
| Input | Free | — | Free |
| Output | Free | — | Free |
Providers
One company serves Nemotron 3 Nano 30B A3B (free). Figures are each provider’s own list price per 1M tokens — your OpenKey price is in the card above.
| Provider | Context | Input /M | Output /M |
|---|---|---|---|
| Nvidiabf16 | 256K | Free | Free |
Provider list from the OpenRouter catalog, Jul 28, 2026. All providers
Benchmarks
Artificial Analysis
- Agentic index
- 2
- Coding index
- 14.4
- Intelligence index
- 14.2
Supported parameters
- include_reasoning
- max_tokens
- reasoning
- seed
- temperature
- tool_choice
- tools
- top_p
Call it
OpenAI-compatible: point your SDK at api.openkey.ai/v1 and use
model nvidia/nemotron-3-nano-30b-a3b:free.
curl https://api.openkey.ai/v1/chat/completions \
-H "Authorization: Bearer ***" \
-H "Content-Type: application/json" \
-d '{
"model": "nvidia/nemotron-3-nano-30b-a3b:free",
"messages": [{"role": "user", "content": "Hello"}]
}'import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.openkey.ai/v1",
api_key=os.environ["OPENKEY_API_KEY"],
)
completion = client.chat.completions.create(
model="nvidia/nemotron-3-nano-30b-a3b:free",
messages=[{"role": "user", "content": "Hello"}],
)
print(completion.choices[0].message.content)import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api.openkey.ai/v1",
apiKey: process.env.OPENKEY_API_KEY,
});
const completion = await client.chat.completions.create({
model: "nvidia/nemotron-3-nano-30b-a3b:free",
messages: [{ role: "user", content: "Hello" }],
});
console.log(completion.choices[0].message.content);About Nemotron 3 Nano 30B A3B (free)
Nemotron 3 Nano 30B A3B is NVIDIA's small mixture-of-experts model, built for developers wiring up agentic AI systems rather than chat products. On Artificial Analysis benchmarks it scores 14.2 on the intelligence index and 14.4 on coding, with an agentic index of 2 — numbers that mark it as a lightweight, low-cost worker model rather than a frontier reasoner. It takes text in and out, supports tool calling, and carries a 256K token context window. Reasoning is optional per request, not forced on every call.
This is the free variant of the paid Nemotron 3 Nano 30B A3B, which itself charges just $0.05 per 1M input tokens and $0.2 per 1M output tokens — so you get the same 256K context window at zero cost. That context window beats 52% of the catalog. Against its own family, it sits below the 550B-parameter Nemotron 3 Ultra (up to $0.5/$2.2 per 1M) in scale but matches it on the free tier, since NVIDIA also ships free versions of Ultra and Super.
Questions
- How much does Nemotron 3 Nano 30B A3B (free) cost via API?
- It's free — $0 per 1M input tokens and $0 per 1M output tokens, both from NVIDIA and via OpenKey (provider price $0 x 1.03 fee = $0). If you need guaranteed throughput or SLAs, the paid version of the same model, nvidia/nemotron-3-nano-30b-a3b, runs $0.05 per 1M input and $0.2 per 1M output.
- What is Nemotron 3 Nano 30B A3B's context window?
- It supports 256,000 tokens of context, enough to hold a few hundred pages of text in a single request. That places it above 52% of models in the OpenKey catalog on context length, despite being one of the smaller Nemotron 3 models by parameter count.
- Is Nemotron 3 Nano 30B A3B (free) actually free to use?
- Yes, this specific version — the ':free' variant — has $0 input and $0 output pricing from NVIDIA, and OpenKey passes that through with no markup since 3% of $0 is still $0. NVIDIA also sells a paid version of the same underlying model at $0.05/$0.2 per 1M tokens for higher-priority access.
- Does Nemotron 3 Nano 30B A3B support tool calling?
- Yes. Its supported parameters include 'tools' and 'tool_choice', so it can call functions during a conversation. It also supports optional reasoning mode, which you can turn on or off per request rather than it being mandatory.
- How does Nemotron 3 Nano 30B A3B compare to Nemotron 3 Ultra?
- Nemotron 3 Nano 30B A3B is the smaller, MoE-based model in NVIDIA's lineup, scoring 14.2 on the Artificial Analysis intelligence index and 14.4 on coding. Nemotron 3 Ultra 550B A55B is the much larger sibling, priced up to $0.5 per 1M input and $2.2 per 1M output tokens on its paid tier — though NVIDIA also offers a free version of Ultra for teams that don't need Nano's smaller footprint.
More from NVIDIA
All NVIDIA models →
NVIDIANemotron 3 Ultra
NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it...
ctx 1M in $0.515 out $2.27
NVIDIAFreeNemotron 3 Ultra (free)
NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it...
ctx 1M in Free out Free
NVIDIAFreeNemotron 3.5 Content Safety (free)
NVIDIA Nemotron 3.5 Content Safety is a compact 4B-parameter multimodal guardrail model from NVIDIA, fine-tuned from Google Gemma-3-4B. It moderates both inputs to and responses from LLMs and VLMs, accepting...
ctx 128K in Free out Free
NVIDIAFreeNemotron 3 Nano Omni (free)
NVIDIA Nemotron™ 3 Nano Omni is a 30B-A3B open multimodal model designed to function as a perception and context sub-agent in enterprise agent systems. It accepts text, image, video, and...
ctx 256K in Free out Free