Requesty
Doubleword

nvidia-nemotron-3-ultra:flex

NVIDIA Nemotron 3 Ultra is NVIDIA's strongest open-weights reasoning model, positioned near GPT-5.4 Mini (xhigh) and ahead of DeepSeek V4-Flash and Qwen3.5-397B-A17B.

Reasoning
Try nvidia-nemotron-3-ultra:flexCompareDocs

Specifications

Context window262K tokens
Max output131K tokens
API typechat
AddedJun 23, 2026
Model IDdoubleword/nvidia-nemotron-3-ultra:flex
Data retentionYes
Used for trainingNo
Provider location🇬🇧 UK, 🌍 Global inference
Privacy policyDoubleword Data Usage 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 Doubleword models overview.

Pricing

Input / 1M
$0.37
Output / 1M
$1.87
Cache write
N/A
Cache read
N/A
Estimated cost
100K input + 10K output$0.0557
1M input + 100K output$0.56
10M input + 1M output$5.57

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 doubleword/nvidia-nemotron-3-ultra: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="doubleword/nvidia-nemotron-3-ultra:flex",
    messages=[
        {"role": "user", "content": "Explain quantum computing in one paragraph."},
    ],
)

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

Other Doubleword models

deepseek-v4-flash:flex
1.0M ctx$0.10/1M
Coding 56
deepseek-v4-flash
1.0M ctx$0.14/1M
Coding 56
deepseek-v4-pro:flex
1.0M ctx$1.31/1M
Coding 59
deepseek-v4-pro
1.0M ctx$1.74/1M
Coding 59
glm-5.2:flex
1.0M ctx$1.05/1M
Coding 69
nvidia-nemotron-3-ultra
262K ctx$0.50/1M

Frequently asked questions

How much does nvidia-nemotron-3-ultra:flex cost?
nvidia-nemotron-3-ultra:flex is priced at $0.37 per million input tokens and $1.87 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-ultra:flex?
nvidia-nemotron-3-ultra:flex has a context window of 262K tokens, with a maximum output of 131K tokens per response. That's roughly 350 words of input you can fit in a single prompt.
What can nvidia-nemotron-3-ultra:flex do?
nvidia-nemotron-3-ultra:flex supports extended reasoning. You can call it through any OpenAI-compatible client by pointing base_url to Requesty.
How do I use nvidia-nemotron-3-ultra: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 "doubleword/nvidia-nemotron-3-ultra:flex". The Quickstart above shows Python, JavaScript and cURL snippets.
Can I run nvidia-nemotron-3-ultra:flex through Requesty?
Yes. nvidia-nemotron-3-ultra:flex runs through Requesty's OpenAI-compatible API, served from Doubleword. You do not host the model yourself: point base_url at Requesty, set the model to "doubleword/nvidia-nemotron-3-ultra:flex", and requests are routed to the upstream provider with automatic failover. The same key gives you 400+ other models too.

Access nvidia-nemotron-3-ultra:flex through Requesty

One API key, 400+ models, OpenAI-compatible. No markup on provider prices, automatic failover, and smart caching built-in.

Get started freeAll Doubleword models