majentik/Nemotron-3-Nano-4B-TurboQuant-MLX-4bit

Nemotron-3-Nano-4B - TurboQuant MLX 4-bit

4-bit weight-quantized MLX version of nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16 with TurboQuant KV-cache quantization. Optimized for Apple Silicon inference via the MLX framework. A good balance between model quality and memory efficiency. The dense hybrid Mamba-2 + Attention architecture supports up to 262K context length.

Approximate model size: ~2.3 GB

Model Specifications

Property	Value
Base Model	nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16
Parameters	4 billion (dense)
Architecture	Hybrid Mamba-2 + Attention (dense)
Context Length	262,144 tokens (262K)
License	NVIDIA Open Model License (commercial use OK)
Weight Quantization	4-bit (~2.3 GB)
KV-Cache Quantization	TurboQuant
Framework	MLX (Apple Silicon)

Quickstart

from mlx_lm import load, generate
from turboquant import TurboQuantCache

model, tokenizer = load("majentik/Nemotron-3-Nano-4B-TurboQuant-MLX-4bit")

prompt = "Explain the theory of relativity."
response = generate(model, tokenizer, prompt=prompt, max_tokens=512)
print(response)

What is TurboQuant?

TurboQuant (arXiv: 2504.19874) is a KV-cache quantization technique that compresses the key-value cache used during autoregressive generation. Combined with 4-bit weight quantization in MLX, this provides a dual compression strategy: smaller model weights plus compressed KV cache for efficient long-context generation.

Key benefits:

• No weight modification -- model weights stay at original precision
• Reduced inference memory -- KV cache is compressed significantly
• Longer context windows -- fit more tokens in the same GPU memory
• Minimal quality loss -- carefully designed quantization preserves generation quality

KV-Cache Quantization Comparison

Method	Prefill Speed	Decode Speed	Memory Savings	Reference
TurboQuant	1x (baseline)	1x (baseline)	High	arXiv: 2504.19874
RotorQuant	5.3x faster	28% faster	High	GitHub

Memory Estimates (Nemotron-3-Nano-4B)

Precision	Approximate Size	MLX Variant
BF16 (original)	~8 GB	--
8-bit quantized	~4 GB	TurboQuant-MLX-8bit
4-bit quantized	~2.3 GB	This model
2-bit quantized	~1.2 GB	TurboQuant-MLX-2bit

Hardware Requirements

This model requires approximately 2.3 GB of unified memory. Recommended hardware:

• Apple M1 (8 GB+)
• Apple M2 (8 GB+)
• Apple M3 (8 GB+)
• Apple M4 (8 GB+)
• Any Apple Silicon Mac with 8 GB+ unified memory

See Also

• nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16 -- Base model
• majentik/Nemotron-3-Nano-4B-TurboQuant -- TurboQuant KV-cache only (transformers)
• majentik/Nemotron-3-Nano-4B-TurboQuant-MLX-8bit -- MLX 8-bit variant
• majentik/Nemotron-3-Nano-4B-TurboQuant-MLX-2bit -- MLX 2-bit variant
• majentik/Nemotron-3-Nano-4B-RotorQuant-MLX-4bit -- RotorQuant MLX 4-bit variant
• TurboQuant Paper (arXiv: 2504.19874)
• MLX Framework