NVIDIA: Nemotron 3 Ultra

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...

Pricing and specs

Providernvidia
Input price$0.600 / 1M tokens
Output price$3.60 / 1M tokens
Context window512k tokens
Max outputNot specified
Modalitiestext
Model IDnvidia/nemotron-3-ultra-550b-a55b

What a typical call costs

A 1,000-token prompt with a 500-token answer on NVIDIA: Nemotron 3 Ultra costs roughly $0.00240. Use the OpenRouter cost calculator to price your own prompt across 400+ models, or the token counter to measure prompt length first.

Run your prompts on NVIDIA: Nemotron 3 Ultra

PromNest stores your prompt library and lets you run any prompt on NVIDIA: Nemotron 3 Ultra with your own OpenRouter key — including side-by-side comparison with other models. Browse 1,000+ ready-made prompt templates to start.

Alternatives from nvidia

NVIDIA: Nemotron 3 Nano 30B A3BNVIDIA: Nemotron 3 Nano 30B A3B (free)NVIDIA: Nemotron 3 Nano Omni (free)NVIDIA: Nemotron 3 SuperNVIDIA: Nemotron 3 Super (free)NVIDIA: Nemotron 3 Ultra (batch)NVIDIA: Nemotron 3 Ultra (free)NVIDIA: Nemotron 3.5 Content Safety (free)

← Back to the full AI model catalog