LLM EXPLORER 60,147 MODELS INDEXED

Mistralai Ministral 3 14B Instruct 2512 MLX MXFP4 by mlx-community

By mlx-community · 526 downloads
HF Repo ↗ Copy URL

Mistralai Ministral 3 14B Instruct 2512 MLX MXFP4 is an open-source language model by mlx-community. Features: 14b LLM, VRAM: 7.2GB, License: apache-2.0, Instruction-Based, LLM Explorer Score: 0.23.

  4-bit   Ar Base model:mistralai/ministral... Base model:quantized:mistralai...   Conversational   De   En   Es   Fr   Instruct   It   Ja   Ko   Mistral-common   Mistral3   Mlx   Nl   Pt   Region:us   Safetensors   Sharded   Tensorflow   Zh
Model Card on HF 🤗: https://huggingface.co/mlx-community/mistralai_Ministral-3-14B-Instruct-2512-MLX-MXFP4 

Mistralai Ministral 3 14B Instruct 2512 MLX MXFP4 Parameters and Internals

LLM NameMistralai Ministral 3 14B Instruct 2512 MLX MXFP4
Repository 🤗https://huggingface.co/mlx-community/mistralai_Ministral-3-14B-Instruct-2512-MLX-MXFP4 
Base Model(s)  Ministral 3 14B Instruct 2512   mistralai/Ministral-3-14B-Instruct-2512
Model Size14b
Required VRAM7.2 GB
Updated2026-07-28
Maintainermlx-community
Model Typemistral3
Instruction-BasedYes
Model Files  5.4 GB: 1-of-2   1.8 GB: 2-of-2
Supported Languagesen fr es de it pt nl zh ja ko ar
Model ArchitectureMistral3ForConditionalGeneration
Licenseapache-2.0
Transformers Version5.0.0.dev0
Tokenizer ClassTokenizersBackend
Padding Token<pad>

Best Alternatives to Mistralai Ministral 3 14B Instruct 2512 MLX MXFP4

Best Alternatives
Context / RAM
Downloads
Likes
Ministral 3 14B Instruct 25120K / 27 GB601
Amaretto 14B0K / 28 GB5183
Note: green Score (e.g. "73.2") means that the model is better than mlx-community/mistralai_Ministral-3-14B-Instruct-2512-MLX-MXFP4.