Nemotron 3 Ultra

Name: Nemotron 3 Ultra
Brand: nvidia
Availability: OnlineOnly

NVIDIA's open reasoning and orchestration model, 55B active parameters out of 550B on a hybrid Transformer-Mamba mixture of experts. Text in, text out, with a context window up to 1M tokens. Built for long-running agent work: orchestration, coding agents, deep research and multi-step reasoning.

Jailbreakall 9levels resisted

Website OpenRouter HuggingFace Blog

Provider: NVIDIA
Released: 2026-06-04
Size: XLARGE
Free API (OpenRouter): Yes, 1 provider (via OpenRouter)
Per 1M tokens: $0.00 in · $0.00 out
Rival’s tab: $0.44; 58/58 receipts survived

Free API endpoint listed

Nemotron 3 Ultra has a free endpoint in Rival’s OpenRouter listings. Free requests have usage limits and depend on provider availability.

Provider data: 9 Sep 2026OpenRouter usage limits

View free endpoint

Get API accessProvider and language code samples

Provider

fromimport openai  OpenAI

client = OpenAI(
"https://openrouter.ai/api/v1"    base_url=,
"$OPENROUTER_API_KEY"    api_key=,
)

response = client.chat.completions.create(
"nvidia/nemotron-3-ultra-550b-a55b"    model=,
"role""user""content""Hello!"    messages=[{: , : }],
)
print(response.choices[0].message.content)

Set OPENROUTER_API_KEY with your OpenRouter API key from openrouter.ai/keys.

Deep analysisPsychometrics, taste index, writing DNA

SubjectiveBench

Taste Index

Across 58 scored outputs

How this is measured →

420.42xThe default

0100headroom →

Taste is judged on an uncapped scale, originality first. The space past 100 is craft today's models rarely reach.

Craft52

Originality37

Plays it safe

share of outputs that are the default answer

24%

Writing DNA

Stylometric Fingerprint

Based on 26 text responses

Tick = global average

Vocabulary Diversity59%

Unique words vs. total words. Higher = richer vocabulary.

Sentence Length15.7 words

Average words per sentence.

Hedging0.26

"Might", "perhaps", "arguably" per 100 words.

Bold Formatting7.7

**Bold** markers per 1,000 characters.

List Usage2.7

Bullet and numbered list items per 1,000 characters.

Section Structure0.75

Markdown headings per 1,000 characters.

Emoji Usage0.13

Emoji per 1,000 characters.

Transitions0.04

"However", "moreover", "furthermore" per 100 words.

Opening Habits

Consistency

79%

Across 26 responses

Favorites

Movie

The Shawshank Redemption

19948.7

Album

The Dark Side of the Moon

Pink Floyd

Book

To Kill a Mockingbird

Harper Lee

City

Kyoto

Japan

Game

The Legend of Zelda: Ocarina of Time

Adventure, Action4.4

Model Responses

53 outputs · $0.44 tracked across 53 receipts

Keep exploring

COMPARE

Nemotron 3 Ultra vs MiniMax M3

Real outputs compared side by side

RANKINGS

Best AI for Complex Reasoning

Sally's brothers, GPT-3 training FLOPs, a document editor syncing on client c...

Compare Nemotron 3 Ultra

Alternatives to Nemotron 3 Ultra

Nemotron 3 Ultra

Free API endpoint listed

Taste Index

Stylometric Fingerprint

Favorites

Model Responses

Nemotron 3 Ultra vs MiniMax M3

Best AI for Complex Reasoning

Compare Nemotron 3 Ultra

Alternatives to Nemotron 3 Ultra

Nemotron 3 Ultra

Free API endpoint listed

Taste Index

Stylometric Fingerprint

Favorites

Model Responses

Nemotron 3 Ultra vs MiniMax M3

Best AI for Complex Reasoning