Devstral Small with VISION support

The vision encoder is taken from Mistral Small, works out-of-the-box with llama.cpp

llama-server -hf ngxson/Devstral-Small-Vision-2505-GGUF

Demo

Output:

GGUF

Model size

23.6B params

Architecture

llama

Hardware compatibility

4-bit

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Base model

Quantized

(35)

this model