- input
- $0.075
- output
- $0.40
- Context
- 262Ktokens
- Released
- 8mo ago
- Size
- 120Bparameters
Cost/1M tokens
Specifications
Model
- Developer
- NVIDIA
- Family
- NemotronNemotron 3
- Type
- Language model
- Released
- March 11, 2026
- Model ID
nvidia/nemotron-3-super-120b-a12b
Capabilities
- Input → output
- Text → Text
- Native context
- 262K tokens
- Max output
- 16K tokens
- Tokenizer
- Other
- Knowledge cutoff
- 2025-06
Weights
- Openness
- Open weights
- Size
- 120B
- Architecture
- nemotron_h_moe
- Families
- nemotron_h_moe
- Hugging Face
- nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-FP8
Sources:models.devOpenRouter
Metadata synced Nov 5, 2026, 6:00 PM UTC