Skip to content

Devstral-Small-2-24B-Instruct-2512-8bit

mlx-community/Devstral-Small-2-24B-Instruct-2512-8bit
8-bit Open on Hugging Face
Parameters
24B
Weights (est.)
≈ 26.6 GB
Mac memory (est.)
32 GB+

Size is estimated from the parameter count and quantisation. Memory allows a fifth extra for context plus 2 GB for the system; long contexts need more.

Run it

With mlx-vlm on an Apple silicon Mac:

pip install mlx-vlmpython -m mlx_vlm.generate --model mlx-community/Devstral-Small-2-24B-Instruct-2512-8bit --image photo.jpg --prompt "Describe this image"

Details

Base model
mistralai/Devstral-Small-2-24B-Instruct-2512 (quantized)
Vendor
Mistral
Task
Vision + text
Quantisation
8-bit
Licence
apache-2.0
Downloads
1.1K in the last 30 days · 6.0K all time · 2 likes
Published
2025-12-09 · updated 2026-07-04 (4mo ago)
Tags
mistral3, mistral-common

Other MLX builds of mistralai/Devstral-Small-2-24B-Instruct-2512

Build Quantisation Weights Mac memory Downloads
Devstral-Small-2-24B-Instruct-2512-OptiQ-4bit 4-bit ≈ 14.1 GB 24 GB+ 1.2K
Devstral-Small-2-24B-Instruct-2512-4bit 4-bit ≈ 15.1 GB 24 GB+ 72K
mistralai_Devstral-Small-2-24B-Instruct-2512-MLX-4Bit 4-bit ≈ 15.1 GB 24 GB+ 1.0K
mistralai_Devstral-Small-2-24B-Instruct-2512-MLX-MXFP4 MXFP4 ≈ 15.1 GB 24 GB+ 916
mistralai_Devstral-Small-2-24B-Instruct-2512-MLX-5Bit 5-bit ≈ 18.0 GB 24 GB+ 799
Devstral-Small-2-24B-Instruct-2512-5bit 5-bit ≈ 18.0 GB 24 GB+ 169
Devstral-Small-2-24B-Instruct-2512-6bit 6-bit ≈ 20.8 GB 32 GB+ 914
mistralai_Devstral-Small-2-24B-Instruct-2512-MLX-6Bit 6-bit ≈ 20.8 GB 32 GB+ 767
mistralai_Devstral-Small-2-24B-Instruct-2512-MLX-8Bit 8-bit ≈ 26.6 GB 32 GB+ 1.3K
Devstral-Small-2-24B-Instruct-2512-8bit (this one) 8-bit ≈ 26.6 GB 32 GB+ 1.1K
mistralai_Devstral-Small-2-24B-Instruct-2512-MLX-BF16 bf16 ≈ 48.0 GB 64 GB+ 641
Devstral-Small-2-24B-Instruct-2512-bf16 bf16 ≈ 48.0 GB 64 GB+ 285