Nemotron 3 Super
- Input / output
- $0.06 / $0.32
- Providers
- 2
- Spread
- 1.88×
- 7 days
- −20.0%
- 30 days
- −23.7%
- Context
- 256K
Providers
| Provider | Input | Output | Blended | vs ref | Context | Uptime 1d | Price since |
|---|---|---|---|---|---|---|---|
| bf16 | $0.06 | $0.32 | $0.125 | ref | |||
| bf16 | $0.075 | $0.40 | $0.156 | +25% | |||
| fp8 | $0.08 | $0.45 | $0.173 | +38% | |||
DeepInfrabf16/priority bf16 | $0.113 | $0.60 | $0.234 | +88% |
- DeepInfraCheapestbf16/flex$0.125$0.06 in · $0.32 outref vs ref256K ctx100% uptimebf16
- $0.156$0.075 in · $0.40 out+25% vs ref256K ctx99.6% uptimebf16
- $0.173$0.08 in · $0.45 out+38% vs ref256K ctx33.9% uptimefp8
- bf16/priority$0.234$0.113 in · $0.60 out+88% vs ref256K ctx91.5% uptimebf16
Blended = (3 × input + output) / 4. Ref $0.125 = median of cheapest 20%.
Price history
USD per 1M tokens · each step is a price change
Prices in USD / 1M tokens.
Changes
About
The cheapest provider for Nemotron 3 Super is DeepInfra at $0.125 per 1M tokens (blended), charging $0.06 per 1M input and $0.32 per 1M output tokens. 2 providers serve it; the most expensive charges 1.88× the typical price. Its best price fell 23.7% over the last 30 days. Context window: 256K tokens. Prices tracked since Oct 3, 2026.
NVIDIA Nemotron 3 Super is a 120B-parameter open hybrid MoE model, activating just 12B parameters for maximum compute efficiency and accuracy in complex multi-agent applications. Built on a hybrid Mamba-Transformer...
- Released
- Mar 11, 2026
- First tracked
- Oct 3, 2026
- Canonical slug
- nvidia/nemotron-3-super-120b-a12b-20230311
- Weights
- nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8