README.md

---
base_model: mistralai/Mistral-Small-4-119B-2603
library_name: mlx
license: apache-2.0
tags:
  - rotorquant
  - kv-cache-quantization
  - mistral
  - moe
  - sparse-moe
  - multimodal
  - quantized
  - mlx
  - 4-bit
  - apple-silicon
  - 256k-context
  - thinking
pipeline_tag: text-generation
---

# Mistral-Small-4-119B-RotorQuant-MLX-4bit

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

This repository provides a 4-bit weight-quantized MLX conversion of [mistralai/Mistral-Small-4-119B-2603](https://huggingface.co/mistralai/Mistral-Small-4-119B-2603) with RotorQuant KV cache quantization support. Designed for efficient inference on Apple Silicon Macs.

## Overview

This model applies two complementary compression techniques:

1. **4-bit weight quantization (MLX)** -- reduces model weights from ~238 GB to ~60 GB
2. **RotorQuant KV cache quantization** -- reduces KV cache from ~32 GB to ~6.5 GB at 256K context

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

## 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 | 4-bit (MLX) |
| KV Cache Quantization | RotorQuant 3-bit |

## Memory Estimates

| Configuration | Weights | KV Cache (256K) | Total |
|---|---|---|---|
| FP16 baseline | ~238 GB | ~32 GB | ~270 GB |
| **This model (4-bit MLX + RotorQuant)** | **~60 GB** | **~6.5 GB** | **~66.5 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

```python
from mlx_lm import load, generate

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

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 RotorQuant?

[RotorQuant](https://github.com/scrya-com/rotorquant) is a rotation-based KV cache quantization method that applies learned rotations before quantizing the key-value cache. Key results on the base model:

- **5.3x faster prefill** compared to unquantized baseline
- **28% faster decode** throughput
- **Perplexity: 6.91** vs 7.07 for unquantized (lower is better)

Because it targets the KV cache rather than weights, it stacks with weight quantization for compounding memory savings.

## See Also

- [mistralai/Mistral-Small-4-119B-2603](https://huggingface.co/mistralai/Mistral-Small-4-119B-2603) -- Base model
- [majentik/Mistral-Small-4-119B-RotorQuant](https://huggingface.co/majentik/Mistral-Small-4-119B-RotorQuant) -- KV cache only (no weight quantization)
- [majentik/Mistral-Small-4-119B-RotorQuant-MLX-2bit](https://huggingface.co/majentik/Mistral-Small-4-119B-RotorQuant-MLX-2bit) -- 2-bit MLX variant
- [majentik/Mistral-Small-4-119B-RotorQuant-MLX-1bit](https://huggingface.co/majentik/Mistral-Small-4-119B-RotorQuant-MLX-1bit) -- 1-bit MLX variant
- [RotorQuant GitHub](https://github.com/scrya-com/rotorquant)