Primary use stated or inferred cautiously from official documentation.
N
NVIDIA / nemotron-3-ultra
NVIDIA Nemotron 3 Ultra
Open 550B-parameter MoE model for orchestration, research, code, and long-duration agents.
01 / SUMMARY
The essential
Maximum input capacity when published by the source.
Documented maximum generation limit.
Verified availability channels.
Primary functional family and verified specialties.
Declared terms for API access, model weights, or self-hosting.
02 / TECHNICAL SHEET
Limits and integration
| API ID | nemotron-3-ultra |
|---|---|
| Model type | Text and reasoning · Code |
| Access model | API and open weights |
| License | NVIDIA open license |
| Deployment | NIM, local, or private cloud |
| Release | 2026 |
| Knowledge cutoff | Not published |
| Entrance | Text |
| Exit | Text |
| Context window | 1,000,000 tokens |
| maximum output | Not published |
| Reasoning | Agentic reasoning |
| Published tools | Features · NVIDIA NIM |
| Structured Outings | Not applicable or not published |
| Batch processing | Not published |
| Prompt cache | Not published |
| Fine-tuning | Published weights, data, and recipes |
| Verified platforms | NVIDIA NIM · Hugging Face |
Enterprise agents, code, and accelerated deployment under your own control.
Its size requires multiple GPUs; compare total cost per task, not just speed per token.
03 / CAPABILITIES
What it can do
Orientation
Enterprise agents, code, and accelerated deployment under your own control.
Context
1.000.000 tokens · Not published
Tools and integration
Functions · NVIDIA NIM
Access
NVIDIA NIM and weights
Input modalities
Text
04 / PRICES
Documented cost
| Concept | Worth | Unit/condition |
|---|---|---|
| Standard input | Not published | Standard API rate |
| Cached input | Not published | Reading reused prefixes |
| Cache write or storage | Not published | The condition varies by provider |
| Standard output | Not published | May include reasoning tokens |
| Batch input | Not published | Asynchronous processing |
| Batch output | Not published | Asynchronous processing |
Consult the primary source: the billing unit depends on the model type and access channel.
i Prices change and may depend on level, region or context length. Check the source before making a decision.
05 / EVALUATIONS
How to read the results
“
A public benchmark provides guidance, but does not replace an evaluation with your data, tools, budget, and error tolerance.
| Benchmark | Result | Metric | Source |
|---|---|---|---|
| Coste en SWE-bench | Up to −30 % | Published cost per completed task | View source ↗ |
Inferama only highlights a “best result” when the metric, test set, configuration, and date allow for an equivalent comparison. The supplier's figures are presented as claims from its own source.
06 / VERSIONS
Chronology
NVIDIA Nemotron 3 Ultra
Version added to Inferama's verified catalog.
ANALYSIS
Related analysis
07 / SOURCES