NVIDIA model product
Nemotron 3 Ultra
Nemotron 3 Ultra is NVIDIA's frontier-scale open model with 550B total / 55B active parameters, built for agentic reasoning, long-context analysis, tool use, and high-stakes RAG.
Updated Aug 12, 2026. Default version: Nemotron 3 Ultra (550B A55B)
LLMBoard score62.9Nemotron 3 Ultra (550B A55B)
Coverage100%23 benchmark families
Context window1MTokens
Official input price$0.50Nvidia API
Nemotron 3 Ultra Specifications
Technical details for the model's default version.
- Version
- Nemotron 3 Ultra (550B A55B)
- Released
- Jun 4, 2026
- Knowledge cutoff
- Sep 30, 2025
- Parameters
- 550B
- Context window
- 1M
- Max output
- 128K
- Inputs
- text
- Outputs
- text
- Open weights
- Yes
- License
- OpenMDW License v1.1
Nemotron 3 Ultra Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Nemotron 3 Ultra Benchmark Results
Benchmark scores for Nemotron 3 Ultra (550B A55B).
| Apex | 84.8% | 01 | 2 | 100.0% | C | |
| IMO-AnswerBench | 92.3% | 01 | 19 | 100.0% | C | |
| OmniScience | 78.7% | 01 | 2 | 100.0% | C | |
| PinchBench | 90.0% | 01 | 4 | 100.0% | C | |
| ProfBench | 56.0% | 01 | 1 | 100.0% | C | |
| RULER | 94.7% | 01 | 4 | 100.0% | C | |
| IFBench | 81.7% | 02 | 29 | 96.4% | C | |
| GDPval | 46.7% | 03 | 3 | 0.0% | C | |
| CritPT | 3.1% | 04 | 4 | 0.0% | C | |
| LiveCodeBench v6 | 89.0% | 04 | 53 | 94.2% | C | |
| LongBench v2 | 61.9% | 04 | 17 | 81.3% | C | |
| MMLU-ProX | 83.0% | 05 | 32 | 87.1% | C | |
| TAU3-Bench | 22.6% | 05 | 5 | 0.0% | C | |
| Multi-Challenge | 63.8% | 06 | 29 | 82.1% | C | |
| WMT24++ | 83.7% | 06 | 23 | 77.3% | C | |
| Finance Agent | 53.7% | 08 | 8 | 0.0% | C | |
| AA-LCR | 65.4% | 09 | 16 | 46.7% | C | |
| MMLU-Pro | 86.8% | 09 | 129 | 93.8% | C | |
| SciCode | 44.6% | 09 | 19 | 55.6% | C | |
| Terminal-Bench 2.1 | 56.4% | 17 | 19 | 11.1% | C | |
| Finance Agent v2 | 37.5% | 20 | 26 | 24.0% | B | |
| SWE-bench Multilingual | 67.7% | 21 | 34 | 39.4% | C | |
| Humanity's Last Exam | 37.4% | 37 | 93 | 60.9% | C | |
| GPQA | 87.0% | 42 | 234 | 82.4% | C | |
| BrowseComp | 44.4% | 49 | 58 | 15.8% | C | |
| SWE-Bench Verified | 70.7% | 57 | 105 | 46.1% | C |
Nemotron 3 Ultra Arena Results
Preference and agent-evaluation results for the default version.
| agent tool hallucination | overall | 33 | 0.0 | N/A | 135.6K | |
| agent bash recovery steps | overall | 43 | -0.2 | N/A | 7.8K | |
| agent praise complaint | overall | 43 | -0.1 | N/A | 878 | |
| agent | overall | 44 | -0.1 | N/A | 152.4K | |
| agent steerability | overall | 46 | -0.2 | N/A | 4.6K | |
| agent task outcome explicit | overall | 46 | -0.2 | N/A | 3.6K | |
| text | overall | 49 | 1445.0 | 10,719 | N/A | |
| text factuality | overall | 95 | 1430.0 | 10,719 | N/A | |
| text style control | overall | 96 | 1426.4 | 10,719 | N/A |
Nemotron 3 Ultra Runtime Performance
Provider-specific output speed and catalog latency for Nemotron 3 Ultra (550B A55B). Runtime does not affect the capability score.
No runtime data
Browse runtime rankingsNo provider-specific speed or latency record is linked to the default version yet.
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Nemotron 3 Ultra Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- $0.50 input, $2.5 output per 1M
- Official provider
- Nvidia
- Lowest third-party
- From $0.10 input, $0.10 output per 1M via routing.run
- Tracked offerings
- 11
| Kenari | nemotron-3-ultra-550b-a55b | global | N/A | N/A | 1M | |
| UnoRouter | nemotron-3-ultra-550b-a55b:free | global | N/A | N/A | 1M | |
| routing.run | nemotron-3-ultra | global | $0.10 | $0.10 | 131.1K | |
| Nvidia | nvidia/nemotron-3-ultra-550b-a55b | global | $0.50 | $2.5 | 1M | |
| NanoGPT | nvidia/nemotron-3-ultra-550b-a55b | global | $0.50 | $2.5 | 1M | |
| LLM Gateway | nemotron-3-ultra-550b | global | $0.50 | $2.5 | 1M | |
| Pioneer | nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 | global | $0.50 | $2.5 | 1M | |
| Kilo Gateway | nvidia/nemotron-3-ultra-550b-a55b | global | $0.50 | $2.2 | 512.3K | |
| Together AI | nvidia/nemotron-3-ultra-550b-a55b | global | $0.60 | $3.6 | 512.3K | |
| Vercel AI Gateway | nvidia/nemotron-3-ultra-550b-a55b | global | $0.60 | $2.4 | 1M | |
| OpenRouter | nvidia/nemotron-3-ultra-550b-a55b | global | $0.60 | $3.6 | 512.3K |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Nemotron 3 Ultra Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
| Nemotron 3 Ultra (550B A55B) | 62.9 | 550B | 1M | 128K | Yes | OpenMDW License v1.1 |
What is Nemotron 3 Ultra?
Key information about Nemotron 3 Ultra and its available data.
Nemotron 3 Ultra is NVIDIA's frontier-scale open model with 550B total / 55B active parameters, built for agentic reasoning, long-context analysis, tool use, and high-stakes RAG.
It uses a hybrid Latent Mixture-of-Experts (LatentMoE) architecture interleaving Mamba-2, MoE, and select Attention layers, with Multi-Token Prediction (MTP) for native speculative decoding, and is pre-trained on ~20T tokens with an NVFP4 recipe.
Reasoning is configurable on/off (plus a medium-effort mode) via the chat template. It supports up to a 1M-token context and 10 languages (English, French, Spanish, Italian, German, Japanese, Hindi, Korean, Brazilian Portuguese, Chinese). 1 license.
Data as of 2026-08-11.
Nemotron 3 Ultra vs nearby models
Open a comparison with the three ranked models immediately above and below this model.
Models similar to Nemotron 3 Ultra
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
NV
Nemotron 3 Super
NVIDIA
NV
Nemotron 3 Nano
NVIDIA
NV
Nemotron Nano 9B
NVIDIA
NV
Llama 3.1 Nemotron Ultra 253B
NVIDIA
NV
Llama 3.3 Nemotron Super 49B
NVIDIA
NV
Llama 3.1 Nemotron Nano 8B
NVIDIA
FAQ
Common questions about Nemotron 3 Ultra.
When was Nemotron 3 Ultra released?
Nemotron 3 Ultra's default version was released on Jun 4, 2026.
How much does Nemotron 3 Ultra cost?
Nemotron 3 Ultra's official API price is $0.50 per million input tokens and $2.5 per million output tokens via Nvidia. The lowest tracked third-party offer starts at $0.10 input and $0.10 output via routing.run.
Who created Nemotron 3 Ultra?
Nemotron 3 Ultra was created by NVIDIA.
What is the context window for Nemotron 3 Ultra?
The default version has a 1M token context window.
Is Nemotron 3 Ultra open weight?
Yes. The default version is marked as open weight under OpenMDW License v1.1.
How many API providers offer Nemotron 3 Ultra?
11 provider offerings are linked to the default version.
What models should I compare Nemotron 3 Ultra with?
Nearby ranked alternatives include Gemini 3 Flash, MiMo V2 Pro, GPT-5.3-Codex.