What are the hardware requirements for Devstral Small 2 24B Instruct 2512 OptiQ 4bit?

Devstral Small 2 24B Instruct 2512 OptiQ 4bit requires approximately 16.5 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Devstral Small 2 24B Instruct 2512 OptiQ 4bit and how large is it?

Devstral Small 2 24B Instruct 2512 OptiQ 4bit is developed by mlx-community, a model with 24b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Devstral Small 2 24B Instruct 2512 OptiQ 4bit?

Devstral Small 2 24B Instruct 2512 OptiQ 4bit is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Devstral Small 2 24B Instruct 2512 OptiQ 4bit by mlx-community — VRAM 16.5GB

Name: Devstral Small 2 24B Instruct 2512 OptiQ 4bit
Author: mlx-community

Devstral Small 2 24B Instruct 2512 OptiQ 4bit is an open-source language model by mlx-community. Features: 24b LLM, VRAM: 16.5GB, License: apache-2.0, Quantized, Instruction-Based, LLM Explorer Score: 0.28.

4-bit 4bit 8bit Apple-silicon Base model:mlx-community/devst... Base model:quantized:mlx-commu... Conversational Devstral Instruct Mistral Mistral3 Mixed-precision Mlx Optiq Quantized Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/mlx-community/Devstral-Small-2-24B-Instruct-2512-OptiQ-4bit

Devstral Small 2 24B Instruct 2512 OptiQ 4bit Parameters and Internals

LLM Name	Devstral Small 2 24B Instruct 2512 OptiQ 4bit
Repository 🤗	https://huggingface.co/mlx-community/Devstral-Small-2-24B-Instruct-2512-OptiQ-4bit
Base Model(s)	...Small 2 24B Instruct 2512 Bf16 mlx-community/Devstral-Small-2-24B-Instruct-2512-bf16
Model Size	24b
Required VRAM	16.5 GB
Updated	2026-08-15
Maintainer	mlx-community
Model Type	mistral3
Instruction-Based	Yes
Model Files	5.4 GB: 1-of-4 5.3 GB: 2-of-4 5.1 GB: 3-of-4 0.7 GB: 4-of-4
Quantization Type	4bit
Model Architecture	Mistral3ForConditionalGeneration
License	apache-2.0
Transformers Version	5.0.0.dev0
Tokenizer Class	TokenizersBackend
Padding Token	<pad>

Rank the Devstral Small 2 24B Instruct 2512 OptiQ 4bit capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to Devstral Small 2 24B Instruct 2512 OptiQ 4bit

Best Alternatives	Context / RAM	Downloads	Likes
...3.2 24B Instruct 2506 Bnb 4bit	0K / 14.5 GB	7430	10
...Instruct 2506 Unsloth Bnb 4bit	0K / 14.6 GB	2107	13
...all 3.2 24B Instruct 2506 8bit	0K / 25 GB	320	4
...l 2 24B Instruct 2512 MLX 8Bit	0K / 25 GB	3915	6
...all 3.2 24B Instruct 2506 4bit	0K / 13.3 GB	682	1
...resolvable Pattern Finetune V1	0K / 48.2 GB	55	0
... 2 24B Instruct 2512 MLX MXFP4	0K / 12.5 GB	1377	4
Devstral Small 2507 GGUF	0K / 0.9 GB	4818	80
...all 3.2 24B Instruct 2506 Bf16	0K / 46.9 GB	128	1

Note: green Score (e.g. "73.2") means that the model is better than mlx-community/Devstral-Small-2-24B-Instruct-2512-OptiQ-4bit.