NVIDIA
Nemotron 3 Ultra 550B A55B (Reasoning)
NVIDIA Nemotron 3 Ultra is an open frontier-reasoning and orchestration model from NVIDIA, with 55B active parameters out of 550B total (MoE). Built on a hybrid Transformer-Mamba mixture-of-experts architecture, it...
NVIDIA model catalogModel Snapshot
ReleaseJun 4, 2026
Context512,288
Overall
Rank #70 across 566
37.8
Coding
Rank #83 across 222
49.3
Blended Price
Rank #193 across 371
$1.18/M
Strength Profile
Percentile score by analysis domain.
* Cost is inverted: lower input, output, and blended prices rank higher.
Benchmark Percentiles
Higher bars mean stronger relative placement.
Model Details
Runtime
Output Speed181.6 tok/s
First Token0.90s
Max Outputn/a
Pricing
Blended$1.18/M
Input$0.675/M
All Benchmarks
| Metric | Domain | Value | Rank |
|---|---|---|---|
| Artificial Analysis Intelligence Index | overall | 37.8 | #70 |
| Artificial Analysis Coding Index | coding | 49.3 | #83 |
| GPQA | reasoning | 86.7% | #67 |
| Humanity's Last Exam | reasoning | 26.6% | #69 |
| SciCode |