All models
NVIDIA
Nemotron 3 Ultra balancedruns locally
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...
Input / 1M
$0.60
USD per million tokens
Output / 1M
$2.40
USD per million tokens
Est. cost / token
$0.00000105
$1.05 per 1M, 3:1 mix
Context window
203K
max output 183K
Major features
Open weightsCachingReasoningJSON schemaTools
Task suitability (heuristic 1–5, not a benchmark)
- coding4
- reasoning4
- math4
- summarization4
- extraction4
- classification4
- chat4
- creative4
- translation4
- vision1
- long context3
- agentic4
Run it locally
Will it run on my machine? →| Quantization | Quality | Memory needed (8K ctx) |
|---|---|---|
| Q2_K | poor | 253 GB |
| Q3_K_M | fair | 295 GB |
| Q4_K_Mrecommended | good | 367 GB |
| Q5_K_M | very good | 426 GB |
| Q6_K | excellent | 490 GB |
| Q8_0 | near-lossless | 632 GB |
| F16 | lossless | 1184 GB |
Estimates. Mixture-of-experts: memory covers all parameters; only the active ones are used per token, so it runs faster than its size suggests.
Price history
| Recorded | Input | Output | Cached | Source |
|---|---|---|---|---|
| 2026-11-05 03:20 | $0.60 | $2.40 | $0.12 | openrouter |
A row is added whenever a sync detects a price change. Vendor pricing source
Pricing (USD per 1M tokens)
- Input
- $0.60
- Output
- $2.40
- Cached input
- $0.12
- Blended (3:1)
- $1.05
Details
- Company
- NVIDIA
- Version
- 3
- Family
- Nemotron
- Released
- 2026-06-04
- Parameters
- 560.5B (55B active)
- Max output
- 183K
- Open weights
- Yes
- Inputs
- text
- Last synced
- 46 min ago