Name: Nemotron 3 Nano 30B A3B API
Brand: NVIDIA

Overview

What Is Nemotron 3 Nano 30B A3B?

30B total parameters, A3B meaning roughly 3 billion are activated per inference pass. NVIDIA built this model from scratch — not fine-tuned on top of someone else's base — training on 25 trillion tokens of text covering code, math, science, general knowledge, and multilingual data.

What makes it genuinely different from most open-weight models in this size class is the architecture. Rather than stacking standard transformer attention layers throughout, NVIDIA combined three distinct layer types — Mamba-2, Mixture-of-Experts, and grouped-query attention — into a hybrid stack that runs significantly faster in practice while handling very long contexts without the usual memory blowup.

Architecture: Hybrid Mamba-Transformer MoE

The architectural backbone of Nemotron 3 Nano is what separates it from straightforward scaled-down transformers. NVIDIA calls it a Hybrid Mamba-Transformer MoE, and understanding each component helps clarify exactly why the model performs the way it does.

Parameter	Value
Total Layers	52
Mamba-2 Layers	23
MoE Layers	23
GQA Attention Layers	6 (2 groups)
Experts per MoE Layer	128 routed + 1 shared
Activated Experts per Token	6
Hidden Dimension	2,688
Activation Function	Squared ReLU (ReLU²)
Normalization	RMSNorm
Context Window	Up to 262,144 tokens (1M supported)

Benchmark Performance

Numbers matter more than marketing, so here's where Nemotron 3 Nano 30B A3B actually lands against its direct open-weight competitors — Qwen3-30B-A3B and GPT-OSS-20B.

Metric	Value
AIME 2025 — Math Reasoning (no tools)	89.1%
Qwen3-30B-A3B (baseline)	85.0%
GPT-OSS-20B	91.7%
AIME 2025 — Math Reasoning (with Python tools)	99.2%
Performance Insight	Near theoretical ceiling with tool integration
Inference Throughput vs Qwen3-30B-A3B (H200, 8K→16K)	3.3× faster
Throughput vs GPT-OSS-20B	2.2× faster (same hardware)

On long-context evaluations (RULER benchmark), Nemotron 3 Nano outperforms both Qwen3-30B-A3B-Instruct-2507 and GPT-OSS-20B across varying context lengths — a direct payoff from the Mamba-2 linear-time design. FP8 quantization retains approximately 99% of BF16 accuracy, meaning the model runs efficiently on 24GB VRAM cards like the RTX 4090 without a significant quality hit.

Capabilities and Use Cases

Nemotron 3 Nano was fine-tuned and RL-trained with production use cases in mind, not benchmark chasing. It handles both reasoning and non-reasoning modes via a flag in the chat template — useful for cutting latency on simpler tasks while retaining full chain-of-thought quality where it matters.

Agentic Systems

Explicitly designed for multi-step agent loops: tool calling, planning, structured output generation, and code execution. The AIME-with-tools score of 99.2% illustrates how well it integrates with external functions.

RAG & Long-Context Processing

A 262K-token context window means large codebases, legal documents, research papers, or extended session histories fit in a single call. Long-range coherence is maintained without the memory overhead of pure-attention models.

Code Generation & Review

Fine-tuned on high-quality code data spanning multiple programming languages. Strong performance on HumanEval and MBPP benchmarks, with a particular edge in software engineering tasks involving agentic tool use.

Mathematical Reasoning

Trained on math-specific synthetic data and RL-reinforced on reasoning tasks. Competitive with much larger models on AIME 2025, MATH-500, and similar benchmarks, especially when Python execution is available.

Custom Fine-Tuning

Open weights, training recipes, SFT datasets, and RL datasets are all released. Deploy on your own infrastructure using vLLM, SGLang, or TensorRT-LLM, with full control over privacy and customization.

API Pricing

$0.065 input
$0.26 output

How It Compares: Nemotron 3 Nano vs the Competition

The MoE-at-30B-with-3B-active category has a few prominent models right now. Here's a grounded comparison:

Model	Active Params	Context	Architecture	Open Weights
Nemotron 3 Nano 30B A3B	~3B	262K	Hybrid Mamba-MoE	✓ Full
Qwen3-30B-A3B	~3B	128K	Standard Transformer MoE	✓ Full
GPT-OSS-20B	~20B	128K	Dense Transformer	✓ Full

The key differentiator for Nemotron 3 Nano is the Mamba-2 integration. While Qwen3-30B-A3B matches it on active parameters and comes close on many benchmarks, the hybrid architecture gives Nemotron a decisive throughput edge on long sequences. For workloads where you're regularly processing 50K+ token inputs — full codebases, lengthy document sets, extended agent histories — the 3.3× throughput advantage is a real operational consideration, not a footnote.

The trade-off is that GPT-OSS-20B edges Nemotron 3 Nano on general knowledge (MMLU) benchmarks. For broad conversational QA, the dense-parameter model has a slight edge. For reasoning with tools, long-context tasks, and agentic workflows, Nemotron 3 Nano's efficiency-per-token math is hard to argue with at this price point.

‍

Uptime and performance

Uptime (7d): 100%

Measured on the AI/ML API gateway across live traffic. Figures are medians unless stated. Last updated 2026-10-12.

Use cases

NVIDIA documents this model as suitable for coding, reasoning, math, and long-context tasks.

It is also intended for AI agents that need high accuracy and efficiency for targeted tasks.

NVIDIA further describes it as a general-purpose reasoning and chat model for instruction-following, chatbots, RAG systems, and other AI-powered applications.

The model is also documented as useful for developers building AI agent systems and tool-using workflows.

How to use Nemotron 3 Nano 30B A3B API

Install any OpenAI-compatible SDK, point it at api.aimlapi.com/v1, and set the model to nvidia/nemotron-3-nano-30b-a3b.

import requests

r = requests.post(
    "https://api.aimlapi.com/v1/chat/completions",
    headers={"Authorization": "Bearer " + AIMLAPI_KEY},
    json={
      "model": "nvidia/nemotron-3-nano-30b-a3b",
      "messages": [
        {
          "role": "user",
          "content": "Hello!"
        }
      ]
    },
)
print(r.json())

const r = await fetch("https://api.aimlapi.com/v1/chat/completions", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.AIMLAPI_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    "model": "nvidia/nemotron-3-nano-30b-a3b",
    "messages": [
      {
        "role": "user",
        "content": "Hello!"
      }
    ]
  }),
});
console.log(await r.json());

curl -X POST https://api.aimlapi.com/v1/chat/completions \
  -H "Authorization: Bearer $AIMLAPI_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"nvidia/nemotron-3-nano-30b-a3b","messages":[{"role":"user","content":"Hello!"}]}'

OpenAI-compatible — swap the base URL and it works with your existing SDK.