LLM EXPLORER 55,709 MODELS INDEXED

Ministral 3 14B Base 2512 8bit by mlx-community

By mlx-community · 216 downloads
HF Repo ↗ Copy URL

Ministral 3 14B Base 2512 8bit is an open-source language model by mlx-community. Features: 14b LLM, VRAM: 15.8GB, Context: 256K, License: apache-2.0, Quantized.

  8-bit   8bit   Ar   Base-model Base model:mistralai/ministral... Base model:quantized:mistralai...   De   Edge   En   Es   Fr   Image-text-to-text   It   Ja   Ko   Ministral   Ministral-3   Mistral-common   Mistral3   Mlx   Multimodal   Nl   Pt   Quantized   Region:us   Safetensors   Sharded   Tensorflow   Vision   Vision-language   Zh
Model Card on HF 🤗: https://huggingface.co/mlx-community/Ministral-3-14B-Base-2512-8bit 

Ministral 3 14B Base 2512 8bit Parameters and Internals

LLM NameMinistral 3 14B Base 2512 8bit
Repository 🤗https://huggingface.co/mlx-community/Ministral-3-14B-Base-2512-8bit 
Base Model(s)  Ministral 3 14B Base 2512   mistralai/Ministral-3-14B-Base-2512
Model Size14b
Required VRAM15.8 GB
Updated2026-08-08
Maintainermlx-community
Model Files  5.3 GB: 1-of-3   5.3 GB: 2-of-3   5.2 GB: 3-of-3
Supported Languagesen fr es de it pt nl zh ja ko ar
Quantization Type8bit
Model ArchitectureMambaSSM
Licenseapache-2.0
Context Length262144
Model Max Length262144
Tokenizer ClassLlamaTokenizerFast
Padding Token<pad>
Vocabulary Size131072

Best Alternatives to Ministral 3 14B Base 2512 8bit

Best Alternatives
Context / RAM
Downloads
Likes
Ministral 3 14B Instruct 2512256K / 15.8 GB110841294
Ministral 3 14B Reasoning 2512256K / 28 GB23495145
...stral 3 14B Instruct 2512 BF16256K / 28 GB2229429
Ministral 3 14B Base 2512256K / 28 GB1059760
Ministral 3 14B Base 2512 Bf16256K / 27.9 GB2251
Note: green Score (e.g. "73.2") means that the model is better than mlx-community/Ministral-3-14B-Base-2512-8bit.