NOUMENOpen app
NOUMEN

nvidia · Open weights

Nvidia Nemotron 3 Ultra 550B

Added Jun 4, 2026Open weights

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Use in the app

Nvidia's Nemotron 3 Ultra 550B A55B model from the Nemotron 3 family. It uses a hybrid Mamba-Transformer MoE architecture. Provider-specific context limits vary, with the longest current route supporting up to 1M context.

  • Reasoning
  • Tools
  • Parallel tools
Input
Text
Output
Text
Discovery tier
Open weights

Pricing

USD per 1M tokens
Input
$0.50
Output
$2.50

Reference prices as published by the model provider. API example

Specifications

Context window
1M
Max output
65.5K
Average output (7 days)
56 tokens
Longer than 5% of 468 models

Side by side

Nvidia Nemotron 3 Ultra 550B compared

Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Nvidia Nemotron 3 Ultra 550BnvidiaThis modelQwen3.5 27B TEEQwenRemove
ProvidernvidiaQwen
Discovery tierOpen weightsOpen weights
AddedJun 4, 2026Mar 13, 2026
Limits
Context window1MBest262.14K
Max output65.54K65.54K
Average output56 tokens1.4K tokens
Pricing, USD per 1M tokens
Input$0.50$0.30Best
Output$2.50$2.40Best
Performance
Throughput111.9 tok/s—
Time to first token1.5 s—
Request success (24h)100.0%—
Benchmark indexes
Intelligence Index22.922.9
Coding Index49.3—
Capabilities
ReasoningYesNo
ToolsYesNo
VisionNoYes
Video inputNoYes
Parallel toolsYesNo
Reasoning effort——

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

  • Nvidia Nemotron 3 Ultra 550B
  • Qwen3.5 27B TEE

Intelligence Index

  • Nvidia Nemotron 3 Ultra 550B22.9
  • Qwen3.5 27B TEE22.9

Nvidia Nemotron 3 Ultra 550B: Better than 56% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • Nvidia Nemotron 3 Ultra 550B: 49.3
    • Qwen3.5 27B TEE: —

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • Nvidia Nemotron 3 Ultra 550B: 3.0%
    • Qwen3.5 27B TEE: —

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • Nvidia Nemotron 3 Ultra 550B: 0.3%
    • Qwen3.5 27B TEE: —

  • AA-Briefcasev1.1

    Agentic knowledge work, rated head to head

    • Nvidia Nemotron 3 Ultra 550B: 876 Elo
    • Qwen3.5 27B TEE: —

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • Nvidia Nemotron 3 Ultra 550B: 1,013 Elo
    • Qwen3.5 27B TEE: —

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • Nvidia Nemotron 3 Ultra 550B: 5.0%
    • Qwen3.5 27B TEE: —

  • AA-LCRv1.1

    Reasoning across long documents

    • Nvidia Nemotron 3 Ultra 550B: 79.3%Top
    • Qwen3.5 27B TEE: 77.7%

  • MLCR-AA

    Long-context reasoning over medical records

    • Nvidia Nemotron 3 Ultra 550B: 5.6%
    • Qwen3.5 27B TEE: —

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • Nvidia Nemotron 3 Ultra 550B: 28.4%Top
    • Qwen3.5 27B TEE: 23.9%

  • GPQA Diamond

    Graduate-level science questions

    • Nvidia Nemotron 3 Ultra 550B: 86.7%Top
    • Qwen3.5 27B TEE: 85.8%

  • IFBench

    Following precise output instructions

    • Nvidia Nemotron 3 Ultra 550B: 81.4%Top
    • Qwen3.5 27B TEE: 75.6%

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • Nvidia Nemotron 3 Ultra 550B: 0.5%
    • Qwen3.5 27B TEE: —

  • SciCode

    Scientific computing in Python

    • Nvidia Nemotron 3 Ultra 550B: 40.3%
    • Qwen3.5 27B TEE: —

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • Nvidia Nemotron 3 Ultra 550B: 36.4%Top
    • Qwen3.5 27B TEE: 32.6%

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • Nvidia Nemotron 3 Ultra 550B: 83.3%
    • Qwen3.5 27B TEE: 93.9%Top

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • Nvidia Nemotron 3 Ultra 550B: 79.3%Top
    • Qwen3.5 27B TEE: 77.7%

Scores from Artificial Analysis, updated Oct 26, 2026. Artificial Analysis

Live routing data

Performance and reliability

Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.

Throughput
111.9 tok/s
Faster than 66% of 254 models
Time to first token
1.5 s
Quicker than 89% of 254 models
Prompt cache hits
11.7%
Request success (24h)
100.0%
Reliable

Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.

Hourly request success · last 72 hoursLowest hour 100.0%
  • Healthy (98% or more)
  • Degraded (90–98%)
  • Down (under 90%)
  • No requests

Developers

Use it through the API

Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.

Read the quickstart
curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "nvidia/nemotron-3-ultra-550b-a55b",
    "messages": [{ "role": "user", "content": "Explain this stack trace." }]
  }'
import os
from openai import OpenAI

client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
    model="nvidia/nemotron-3-ultra-550b-a55b",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

nvidia

Other nvidia models, starting with the same family.