---

language:

• en
• multilingual

tags:

• mistral
• ministral
• instruct
• mxfp4
• gguf
• vision
• multimodal
• coding

license: apache-2.0

base_model: mistralai/Ministral-3-8B-Instruct-2512

---

Ministral 3 8B Instruct-2512 — MXFP4 GGUF

MXFP4 quantization of mistralai/Ministral-3-8B-Instruct-2512, an 8B parameter instruct model from Mistral AI.

About the Model

Ministral 3 8B Instruct is a dense transformer with 40 layers, 4096 hidden dimension, and 24-layer Pixtral ViT vision encoder. It supports:

• Code generation and debugging across multiple languages
• Vision understanding via multimodal image input
• Tool calling with native function calling support
• 262K context window with YaRN scaling

Quantization

This GGUF was quantized from Mistral's official BF16 GGUF using llama.cpp (build 537). The BF16 weights were dequantized to F32 during quantization, then quantized to MXFP4 format.

MXFP4 (Microscaling FP4) uses block-wise quantization with shared exponents per block.

Files

| File | Size | Description |

|------|------|-------------|

| ministral-3-8b-instruct-2512-mxfp4.gguf | ~3.5 GB | MXFP4 quantized model weights |

| mmproj-ministral-3-8b-instruct-2512-f16.gguf | ~693 MB | Vision projector (F16, unquantized) |

Usage

llama.cpp

# Server mode with OpenAI-compatible API
llama-server \
  -m ministral-3-8b-instruct-2512-mxfp4.gguf \
  --mmproj mmproj-ministral-3-8b-instruct-2512-f16.gguf \
  -ngl 99 \
  --host 0.0.0.0 \
  --port 8080

LM Studio

1. Download both files from this repository
2. Load the main GGUF file in LM Studio
3. Load the mmproj file for vision support
4. Set GPU offload layers to maximum

Architecture

• Parameters: 8B (dense transformer)
• Layers: 40
• Hidden dimension: 4096
• Attention heads: 32 (8 KV heads for GQA)
• Vision encoder: 24-layer Pixtral ViT
• Context: 262K (YaRN scaled)
• Vocabulary: Mistral Tekken tokenizer

Hardware Requirements

• Minimum: 6 GB VRAM for text-only
• Recommended: 8 GB VRAM for full GPU offload

License

Apache 2.0 — same as the base model.