Model Library
NVIDIAReleased June 4, 2026Nemotron 3 Ultra
NVIDIA · Released June 4, 2026
Intelligence score
38
At a glance
- Intelligence
- 38
- Speed
- 250.2tokens/sec
- Latency
- 1.16stime to first token
- Input price
- $0.68per 1M tokens
- Output price
- $2.67per 1M tokens
- Context window
- 260Ktokens
- Parameters
- 550billions
Summary
Comparison summary
Nemotron 3 Ultra 550B A55B (Reasoning) scores 38 on the Artificial Analysis Intelligence Index, placing it well above average among other open weight models of similar size (median: 25). Nemotron 3 Ultra 550B A55B (Reasoning) generates output at 250.2 tokens per second (based on the median across providers serving the model), which is well above average compared to other open weight models of similar size (median: 61.7 t/s). Nemotron 3 Ultra 550B A55B (Reasoning) costs $0.68 per 1M input tokens (somewhat higher than average, median: $0.59) and $2.67 per 1M output tokens (somewhat higher than average, median: $2.20), based on the median across providers serving the model.
Specifications
Technical specifications
| Model name | Nemotron 3 Ultra |
|---|---|
| Provider | NVIDIA |
| Release date | June 4, 2026 |
| Intelligence score | 38 |
| Speed | 250.2 tokens/sec |
| Latency (TTFT) | 1.16s |
| Input price | $0.68 / 1M tokens |
| Output price | $2.67 / 1M tokens |
| Context window | 260,000 tokens |
| Parameters | 550B |
| Reasoning | Yes |
| Input modalities | text |
| Output modalities | — |
| Open weights | Yes |
| License | OpenMDW |
| API availability | — |
| Source | Model Intelligence legacy profile: nvidia-nemotron-3-ultra-550b-a55b |
Pricing
Token prices
Price per 1M tokens
Nemotron 3 Ultra, as listed in the catalog record
- Input$0.68
- Output$2.67
| Token type | Price per 1M |
|---|---|
| Input | $0.68 |
| Output | $2.67 |
Model Intelligence legacy profile: nvidia-nemotron-3-ultra-550b-a55b
FAQ · 21
Frequently asked questions
Nemotron 3 Ultra 550B A55B (Reasoning) was released on June 4, 2026.
Nemotron 3 Ultra 550B A55B (Reasoning) was created by NVIDIA.
Nemotron 3 Ultra 550B A55B (Reasoning) scores 38 on the Artificial Analysis Intelligence Index, placing it well above average among other open weight models of similar size (median: 25).
Nemotron 3 Ultra 550B A55B (Reasoning) generates output at 250.2 tokens per second (based on the median across providers serving the model), which is well above average compared to other open weight models of similar size (median: 61.7 t/s).
Nemotron 3 Ultra 550B A55B (Reasoning) has a time to first token (TTFT) of 1.16s (based on the median across providers serving the model), which is very competitive compared to other open weight models of similar size (median: 2.01s).
Nemotron 3 Ultra 550B A55B (Reasoning) costs $0.68 per 1M input tokens (somewhat higher than average, median: $0.59) and $2.67 per 1M output tokens (somewhat higher than average, median: $2.20), based on the median across providers serving the model.
Nemotron 3 Ultra 550B A55B (Reasoning) costs $0.68 per 1M input tokens and $2.67 per 1M output tokens (based on the median across providers serving the model). For a blended rate (7:2:1 cache hit/input/output ratio), this is $0.58 per 1M tokens. Pricing may vary by provider.
When evaluated on the Intelligence Index, Nemotron 3 Ultra 550B A55B (Reasoning) generated 110M output tokens, which is somewhat higher than average compared to other open weight models of similar size (median: 92M).
Yes, Nemotron 3 Ultra 550B A55B (Reasoning) is a reasoning model. It uses extended thinking or chain-of-thought reasoning to work through complex problems before providing an answer.
Nemotron 3 Ultra 550B A55B (Reasoning) supports text input.
Nemotron 3 Ultra 550B A55B (Reasoning) supports text output.
No, Nemotron 3 Ultra 550B A55B (Reasoning) does not support image input. It can only process text.
No, Nemotron 3 Ultra 550B A55B (Reasoning) is not multimodal. It only supports text input.
Nemotron 3 Ultra 550B A55B (Reasoning) has a context window of 260k tokens. This determines how much text and conversation history the model can process in a single request.
Yes, Nemotron 3 Ultra 550B A55B (Reasoning) is open weights. The model weights are publicly available.
Nemotron 3 Ultra 550B A55B (Reasoning) has 550 billion parameters (55 billion active).
Nemotron 3 Ultra 550B A55B (Reasoning) is a Mixture of Experts (MoE) model with 550 billion total parameters, but only 55 billion active parameters are used during inference.
Nemotron 3 Ultra 550B A55B (Reasoning) is released under the OpenMDW license. This license allows commercial use.
Nemotron 3 Ultra 550B A55B (Reasoning) achieves a score of 38 on the Artificial Analysis Intelligence Index. This composite benchmark evaluates models across reasoning, knowledge, mathematics, and coding.
Yes, Nemotron 3 Ultra 550B A55B (Reasoning) is available via API through 8 providers.
Nemotron 3 Ultra 550B A55B (Reasoning) is available through 8 API providers.