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: Nemotron 3 Ultra (free) — NVIDIA. Context window 1,000,000 tokens. API pricing (per 1M tokens): Input $0, Output $0. License: Open-Weights. GPQA Diamond 85.4%.
| Provider | NVIDIA |
|---|---|
| API model ID | nvidia/nemotron-3-ultra-550b-a55b:free |
| Tier | Frontier |
| Context window | 1,000,000 tokens |
| Max output | 65,536 tokens |
| License | Open-Weights |
| API pricing (per 1M tokens) · Input | $0 |
| API pricing (per 1M tokens) · Output | $0 |
| GPQA Diamond | 85.4% |
LMArena (CC-BY 4.0, best reasoning setting) · GPQA Diamond: Epoch AI (CC-BY 4.0)
Compare in the full catalog