NVIDIA

Nemotron 3 Ultra 550BOpen

NVIDIA’s 550B open MoE built for enterprise reasoning pipelines.

#561released Jun 3, 2026·Open Weights·550B (55B active MoE)
text512K ctx
API providersGet API Key →Leaderboard
Intelligence
19.0 #561
AAQI 38.0
Output Speed
142 t/s
TTFT 2.15 s · ITL 7.0 ms/tok
Blended price (3:1)
$1.137 /1M
In $0.6/M · Out $2.75/M
Context
512K
≈ 1,280 A4 pages · max out 8,192

Quality Domain Radar

6-Axis Frontier Evaluation vs Category Average & Flagship

Nemotron 3 Ultra 550B
Claude Mythos Preview (#1)
Average (68)
Coding (19.1)Mathematics (15)Hard Science (GPQA) (33.4)Agentic & Tool Use (15)Knowledge (MMLU) (37.5)Vision / Multimodal (25)

Radar axes are approximated from category scores where direct benchmark measurements are missing — see Methodology.

API deployment

Compare hosts & latency

Throughput, TTFT and cache pricing across Groq, Cerebras, Together, DeepInfra, Fireworks, Bedrock and Azure.

GroqCerebrasTogether AIDeepInfraFireworks AIAWS BedrockAzure AI
View providers matrix →

Scores by Source

Arena.ai ↗
1426.0
#96 · 10.6k votes · ±7
Artificial Analysis ↗
38.0
Intelligence Index
LLM Stats ↗
—
Composite score
LLMBase ↗
37.8
Overall index

Category Scores

Value96.8Long Context93.4Knowledge37.5Reasoning33.4Speed30.2Coding19.1Agentic0.0

Benchmark Results

GPQA Diamond87%MMLU-Pro86.8%τ²-Bench Retail83.3%
Ifbench81.4%
LiveCodeBench67%SciCode39.9%Terminal-Bench Hard36.4%Humanity’s Last Exam26.6%

Model Details

Release date
Jun 3, 2026
Parameters
550B (55B active MoE)
License
Open Weights
Context Window
512,000 tokens
Input Price
$0.6/M
Output Price
$2.75/M
Output Speed
142 t/s
Time to First Token
2.15 s

Estimated monthly bill

Estimated monthly spend
$0.00
Calculating savings vs GPT-4o...

Cite & embed

GitHub Markdown badge
Loading badge markdown...
BibTeX
@misc{llmpodium2026nemotron3ultra,
  title = {Benchmark Analysis and Intelligence Evaluation of Nemotron 3 Ultra 550B},
  author = {LLMPodium Research},
  year = {2026},
  url = {https://llmpodium.com/models/nemotron-3-ultra}
}

Embed Score Badge

Add a live Podium Score badge to your GitHub or Hugging Face README — it links back to the live ranking.

Preview
Nemotron 3 Ultra 550B LLMPodium Score
Markdown
Loading badge snippet...
HTML
Loading badge snippet...

More from NVIDIA

Nvidia Nemotron 3 Ultra 550b A55b Nvfp4
Unknown · 50 t/s
71.0
Nvidia Llama 3.3 Nemotron Super 49b V1.5
Unknown · 50 t/s
67.0
Nvidia Nemotron 3 Nano 30b A3b Bf16
Unknown · 50 t/s
66.0
Nemotron 4 340b Instruct
Unknown · 50 t/s
64.0
Nvidia Nemotron 3 Ultra 550b A55b
Proprietary · 150 t/s
38.3
Nvidia Nemotron 3 Super 120b A12b
Proprietary · 150 t/s
32.0

Alternative Models

Closest alternatives to Nemotron 3 Ultra 550B based on capabilities, pricing, and benchmark scores.

Nvidia Nemotron 3 Ultra 550b A55b Nvfp4
NVIDIA · Unknown
71.0
Nvidia Llama 3.3 Nemotron Super 49b V1.5
NVIDIA · Unknown
67.0
Nvidia Nemotron 3 Nano 30b A3b Bf16
NVIDIA · Unknown
66.0
Nemotron 4 340b Instruct
NVIDIA · Unknown
64.0
Nvidia Nemotron 3 Ultra 550b A55b
NVIDIA · Proprietary
38.3
Nvidia Nemotron 3 Super 120b A12b
NVIDIA · Proprietary
32.0
0 / 4
Compare Now →