Nemotron-3 Nano (30B, MoE Reasoning)
NVIDIA's Nemotron-3 Nano Omni — 30B-parameter hybrid Mamba/Transformer Mixture-of-Experts (~3B active per token) tuned for reasoning. Native function-calling, 128K context, multimodal (text + image + audio + video).
At a glance
- Parameters
- 30B
- Quantization
- Q4_K_M
- Context window
- 131K tokens
- Approx. size
- 24.5 GB
- Engines
- llama.cpp, MLX
- License
- NVIDIA Open Model License (custom-restricted)
- Version
- v1.0.0 , released 2026-05-17
- Family
- nemotron
Tuning
Defaults a gezel client applies out of the box, from real evaluation runs against this exact quantization.
- sampling.temperature
0.2- sampling.topK
1- reasoning.thinkingBudget
8192- reasoning.enableThinking
true- profiles
- thinking-general, thinking-coding, thinking-precise, instruct, creative
Model behaviors
Client-side behavior modules the catalog enables for this model (reasoning-tag handling, fabrication detection, prompt shaping).
- reasoning.strip-think-tags
- prompt.private-reasoning-guidance
- fabrication.detect-past-tense-no-tools
- turn.preamble-folding
Sources
- llama.cpp (GGUF)
- lmstudio-community/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-GGUF ·
Nemotron-3-Nano-Omni-30B-A3B-Reasoning-Q4_K_M.gguf· Q4_K_M - MLX
- mlx-community/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-nvfp4
- Upstream
- https://huggingface.co/nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning
- Manifest
- View on GitHub
Tags
- nvidia
- reasoning
- moe
- hybrid
- tools
- multimodal
- vision
- audio
- video