Back to Calculator
Nemotron 3 Super 120B
HotNVIDIA's hybrid Mamba-Transformer MoE with 1M context, optimized for collaborative agents and high-volume enterprise workflows
Model Specifications
ArchitectureTEXT
Parameters120B
Familynemotron
VRAM (Q4)60.0GB
Mixture of ExpertsActive inference parameters: 12B.
Hybrid LatentMoE with Mamba-2 and MTP layers. Supports 1M context
Estimated Quantization Sizes
| Format | Precision | Est. VRAM | Recommendation |
|---|---|---|---|
| FP16 / BF16 | 16-bit | 240.0 GB | Uncompressed Base |
| Q8_0High | 8-bit | 120.0 GB | Near Lossless |
| Q6_K | 6-bit | 90.0 GB | Excellent Balance |
| Q4_K_MPopular | 4-bit | 60.0 GB | Standard Use |
Share this Model
Send this model's specs directly to your community.
Similar Models
Related Guides
How much VRAM do you really need?
A complete breakdown of quantization levels and VRAM overhead for running local models.
Best GPUs for Machine Learning in 2026
Comparing NVIDIA and AMD options for the best speed-to-dollar ratio.
GGUF vs EXL2 vs AWQ
Understanding local AI formats and which one to pick for your specific hardware.