All Models

Nemotron 3 Super 120B

nemotronReasoningTool CallingOpen WeightsStructured Output

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...

Providers4
ReleasedMar 1, 2026
Input Modalitiestext
Output Modalitiestext
Tarsk Usecoding

Benchmarks

12.8
Intelligence Index
37.7
Coding Index

Available Providers (4)

ProviderModel IDInput CostOutput CostContextMax OutputDocs
NanoGPTnvidia/nemotron-3-super-120b-a12b$0.05/MTok$0.25/MTok262.1K16.4K
Perplexity Agentnvidia/nemotron-3-super-120b-a12b$0.25/MTok$2.50/MTok1M32K
DigitalOceannvidia-nemotron-3-super-120b$0.30/MTok$0.65/MTok1M32.8K
Cloudflare Workers AI@cf/nvidia/nemotron-3-120b-a12b$0.50/MTok$1.50/MTok256K256K

Capabilities

Reasoning
Tool Calling
Attachments
Open Weights
Structured Output