LLM EXPLORER 58,133 MODELS INDEXED

Devstral 2 123B Instruct 2512 4bit by mlx-community

By mlx-community · 230 downloads
HF Repo ↗ Copy URL

Devstral 2 123B Instruct 2512 4bit is an open-source language model by mlx-community. Features: 123b LLM, VRAM: 70GB, Context: 256K, License: other, Quantized, Instruction-Based, LLM Explorer Score: 0.22.

  4-bit   4bit Base model:mistralai/devstral-... Base model:quantized:mistralai...   Conversational   Instruct   Ministral3   Mistral-common   Mlx   Quantized   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/mlx-community/Devstral-2-123B-Instruct-2512-4bit 

Devstral 2 123B Instruct 2512 4bit Parameters and Internals

LLM NameDevstral 2 123B Instruct 2512 4bit
Repository 🤗https://huggingface.co/mlx-community/Devstral-2-123B-Instruct-2512-4bit 
Base Model(s)  Devstral 2 123B Instruct 2512   mistralai/Devstral-2-123B-Instruct-2512
Model Size123b
Required VRAM70 GB
Updated2026-08-10
Maintainermlx-community
Model Typeministral3
Instruction-BasedYes
Model Files  5.4 GB: 1-of-14   5.3 GB: 2-of-14   5.2 GB: 3-of-14   5.4 GB: 4-of-14   5.3 GB: 5-of-14   5.2 GB: 6-of-14   5.2 GB: 7-of-14   5.2 GB: 8-of-14   5.4 GB: 9-of-14   5.3 GB: 10-of-14   5.2 GB: 11-of-14   5.2 GB: 12-of-14   5.2 GB: 13-of-14   1.5 GB: 14-of-14
Quantization Type4bit
Model ArchitectureMinistral3ForCausalLM
Licenseother
Context Length262144
Model Max Length262144
Transformers Version5.0.0.dev0
Tokenizer ClassTokenizersBackend
Padding Token<pad>
Vocabulary Size131072

Best Alternatives to Devstral 2 123B Instruct 2512 4bit

Best Alternatives
Context / RAM
Downloads
Likes
...tral 2 123B Instruct 2512 8bit256K / 133 GB1161
Devstral 2 123B Instruct 2512256K / 196.3 GB57919329
Devstral 2 123B Instruct 2512256K / 69.6 GB110
Note: green Score (e.g. "73.2") means that the model is better than mlx-community/Devstral-2-123B-Instruct-2512-4bit.