Requesty
DeepInfra Inc. logo
DeepInfra Inc.

nvidia/Nemotron-3-Nano-30B-A3B:flex

NVIDIA Nemotron 3 Nano is an open small reasoning model optimized for fast, cost-efficient inference in agentic and production workloads. Built with a hybrid Mixture-of-Experts (MoE) and Mamba-Transformer architecture, it delivers strong multi-step reasoning, high token throughput, stable latency with predictable cost, and efficient deployment for agent-based systems. Designed for real-world AI systems where reasoning can generate significantly more tokens per prompt, Nemotron Nano reduces compu

ReasoningTool calling
CompareDocs

Specifications

Context window262K tokens
Max outputN/A
API typechat
AddedAug 14, 2026
Model ID
Data retentionNo
Used for trainingNo
Provider location🇺🇸 US
Privacy policyDeepInfra Privacy Policy→

Benchmarks

Benchmarks haven't been published yet for this exact variant.

Some variants (region-specific deployments, highspeed tiers) share benchmarks with their base model. Check the base model page or the DeepInfra Inc. models overview.

Pricing

Prices updated August 15, 2026
Input / 1M
$0.04
Output / 1M
$0.16
Cache write
N/A
Cache read
N/A
Estimated cost
100K input + 10K output$0.0056
1M input + 100K output$0.0560
10M input + 1M output$0.56

Requesty charges exactly what the upstream provider charges, no markup, no per-request fees. Prompt caching and smart routing can reduce effective cost by 30-80%.

Quickstart

Drop-in compatible with the OpenAI SDK. Change the base URL, swap in your Requesty API key, and set the model to deepinfra/nvidia/Nemotron-3-Nano-30B-A3B:flex.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
from openai import OpenAI

client = OpenAI(
    api_key="YOUR_REQUESTY_API_KEY",
    base_url="https://router.requesty.ai/v1",
)

response = client.chat.completions.create(
    model="deepinfra/nvidia/Nemotron-3-Nano-30B-A3B:flex",
    messages=[
        {"role": "user", "content": "Explain quantum computing in one paragraph."},
    ],
)

print(response.choices[0].message.content)

Other DeepInfra Inc. models

Qwen/Qwen3.5-397B-A17B:flex
262K ctx$0.36/1M
Coding 48
XiaomiMiMo/MiMo-V2.5-Pro:flex
1.0M ctx$0.80/1M
Coding 60
deepseek-v4-flash-0731:flex
1.0M ctx$0.07/1M
Coding 69
Qwen/Qwen3-Coder-480B-A35B-Instruct-Turbo:flex
262K ctx$0.24/1M
deepseek-ai/DeepSeek-V4-Flash:flex
1.0M ctx$0.07/1M
Coding 69
meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo:flex
131K ctx$0.02/1M

Frequently asked questions

How much does nvidia/Nemotron-3-Nano-30B-A3B:flex cost?
nvidia/Nemotron-3-Nano-30B-A3B:flex is priced at $0.04 per million input tokens and $0.16 per million output tokens when accessed via Requesty. Requesty charges exactly what the upstream provider charges, we don't add markup.
What is the context window of nvidia/Nemotron-3-Nano-30B-A3B:flex?
nvidia/Nemotron-3-Nano-30B-A3B:flex has a context window of 262K tokens. That's roughly 350 words of input you can fit in a single prompt.
What can nvidia/Nemotron-3-Nano-30B-A3B:flex do?
nvidia/Nemotron-3-Nano-30B-A3B:flex supports tool calling, extended reasoning. You can call it through any OpenAI-compatible client by pointing base_url to Requesty.
How do I use nvidia/Nemotron-3-Nano-30B-A3B:flex with the OpenAI SDK?
Install the OpenAI SDK, set base_url to "https://router.requesty.ai/v1", set your API key to your Requesty key, and set the model to "deepinfra/nvidia/Nemotron-3-Nano-30B-A3B:flex". The Quickstart above shows Python, JavaScript and cURL snippets.
Can I run nvidia/Nemotron-3-Nano-30B-A3B:flex through Requesty?
Yes. nvidia/Nemotron-3-Nano-30B-A3B:flex runs through Requesty's OpenAI-compatible API, served from DeepInfra Inc.. You do not host the model yourself: point base_url at Requesty, set the model to "deepinfra/nvidia/Nemotron-3-Nano-30B-A3B:flex", and requests are routed to the upstream provider with automatic failover. The same key gives you 600+ other models too.

Access nvidia/Nemotron-3-Nano-30B-A3B:flex through Requesty

One API key, 600+ models, OpenAI-compatible. No markup on provider prices, automatic failover, and smart caching built-in.

All DeepInfra Inc. models