Mistral-Small-4-119B-TurboQuant-MLX-8bit

Dual compression: 8-bit MLX weight quantization + TurboQuant KV cache quantization for Mistral Small 4 on Apple Silicon.

This repository provides an 8-bit weight-quantized MLX conversion of mistralai/Mistral-Small-4-119B-2603 with TurboQuant KV cache quantization support. Designed for efficient inference on Apple Silicon Macs.

Approximate model size: ~120 GB

Overview

This model applies two complementary compression techniques:

1. 8-bit weight quantization (MLX) -- reduces model weights from ~238 GB to ~120 GB
2. TurboQuant KV cache quantization -- reduces KV cache from ~32 GB to ~8 GB at 256K context

Together, these make it feasible to run a 119B-parameter MoE model on high-memory Apple Silicon machines.

Model Specs

Property	Value
Base Model	Mistral Small 4 (March 2026)
Total Parameters	119B
Active Parameters	6.5B per token (Sparse MoE)
Architecture	Sparse MoE -- 128 experts, 4 active per token
Context Length	256K tokens
Modality	Text + Images (multimodal)
Capabilities	Thinking / reasoning, tool use, multilingual
License	Apache 2.0
Weight Quantization	8-bit (MLX)
KV Cache Quantization	TurboQuant 4-bit

Memory Estimates

Configuration	Weights	KV Cache (256K)	Total
FP16 baseline	~238 GB	~32 GB	~270 GB
This model (8-bit MLX + TurboQuant)	~120 GB	~8 GB	~128 GB

Note: This is a Sparse MoE model -- only 6.5B parameters are active per token, so inference is fast despite the 119B total parameter count.

Quickstart

from mlx_lm import load, generate

model, tokenizer = load("majentik/Mistral-Small-4-119B-TurboQuant-MLX-8bit")

prompt = "Explain sparse mixture-of-experts architectures."
messages = [{"role": "user", "content": prompt}]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)

response = generate(model, tokenizer, prompt=text, max_tokens=512)
print(response)

What is TurboQuant?

TurboQuant (arXiv: 2504.19874) is a KV cache quantization method that compresses the key-value cache used during autoregressive generation. It supports 4-bit (default) and 2-bit (aggressive) modes. Because it targets the KV cache rather than weights, it stacks with weight quantization for compounding memory savings.

KV-Cache Quantization Comparison

Method	Prefill Speed	Decode Speed	Memory Savings	Reference
TurboQuant	Baseline	Baseline	High	arXiv: 2504.19874
RotorQuant	5.3x faster	28% faster	High	GitHub

Hardware Requirements

This model requires approximately 128 GB total memory at 256K context. Recommended hardware:

• Apple M2 Ultra (192 GB+)
• Apple M3/M4 Ultra (192 GB+)
• Mac Pro

See Also

• mistralai/Mistral-Small-4-119B-2603 -- Base model
• majentik/Mistral-Small-4-119B-TurboQuant -- KV cache only (no weight quantization)
• majentik/Mistral-Small-4-119B-TurboQuant-MLX-4bit -- 4-bit MLX variant
• majentik/Mistral-Small-4-119B-TurboQuant-MLX-2bit -- 2-bit MLX variant
• majentik/Mistral-Small-4-119B-RotorQuant-MLX-8bit -- RotorQuant MLX 8-bit variant
• TurboQuant Paper (arXiv: 2504.19874)
• MLX Framework