NVIDIA: Nemotron 3 Ultra
nvidia/nemotron-3-ultra-550b-a55b
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...
Modalities
In / Out Price
$0.50 / $2.20 per 1M
Context
262K
Released
Jun 4, 2026
Knowledge Cutoff
—
Pricing
What Reroute charges per unit. No markup on inference — you pay the carrier's price, metered from the carrier's own usage numbers.
Input tokens
$0.50 / M
Output tokens
$2.20 / M
Cache read
$0.10 / M
By provider
Each carrier sets its own price. The router weighs these against uptime when it picks one.
| Provider | Input /M | Output /M | Cache read /M | Cache write /M | Discount |
|---|---|---|---|---|---|
 DeepInfra | $0.50 | $2.20 | $0.10 | — | — |
 BaseTen | $0.60 | $2.40 | $0.12 | — | — |
| BaseTen | $0.60 | $2.40 | $0.12 | — | — |
 Venice | $0.625 | $3.13 | $0.188 | — | — |