Model specification · NVIDIA

Nemotron 3 Super

nvidia/nemotron-3-super-120b-a12b

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...

SpecificationsReported values
Context window
1M
Maximum output
16K
Architecture
text->text
Tokenizer
Other
Knowledge cutoff
Moderated
No
Input modalities
text
Output modalities
text
Full pricingUSD where reported
PriceStandardfree
Input$0.085 / Mfree / M
Output$0.400 / Mfree / M
Benchmark indexesReported indexes
Intelligence
25.4
Coding
37.7
Agentic
8.7
Design Arena Elo
Example workload costsStandard pricing only
WorkloadInput / outputCache shareEstimated cost
Quick check10K / 1K × 10%$0.00125
Long analysis100K / 4K × 150%$0.010
Workday20K / 2K × 1,00030%$2.50
More from NVIDIAAll NVIDIA models
ModelContextInput / MOutput / MIntel.
Nemotron 3 Ultra512K$0.600$3.6037.8
Nemotron 3 Nano 30B A3B262K$0.050$0.20014.2