SPECS
| context window | 262,144 tokens |
|---|---|
| max output | 182,520 tokens |
| price in / out | $0.6 / $2.4 per 1M tokens |
| weights | open-weights |
openrouter.ai/api/v1/models (as of 2026-11-12)
MODEL / NEMOTRON-ULTRA
NVIDIA's open reasoning model.
modelfree tier
SPECS
| context window | 262,144 tokens |
|---|---|
| max output | 182,520 tokens |
| price in / out | $0.6 / $2.4 per 1M tokens |
| weights | open-weights |
openrouter.ai/api/v1/models (as of 2026-11-12)
PUBLISHED BENCHMARKS
| aaIndex | 11.8 |
|---|---|
| aaCodingIndex | 49.3 |
| aaAgenticIndex | 20.1 |
NVIDIA published · 2025
HUGGING FACE · LIVE
| pipeline | text-generation |
|---|---|
| parameters | 560.5B |
| downloads | 1,050,000 |
| likes | 384 |
| last modified | 2026-08-24 |
nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 · observed 2026-11-12 16:53:35
MACHINE-READABLE
GET /api/v1/resources/nemotron-ultraOpen JSON ↗