Converted from mistralai/Ministral-3-14B-Instruct-2512-BF16 using mlx_vlm.convert with --quant-predicate mixed_4_8 on 2026-03-05, 19:48.

Downloads last month
78
Safetensors
Model size
3B params
Tensor type
BF16
·
U32
·
MLX
Hardware compatibility
Log In to add your hardware

4-bit

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

Model tree for dumtjul/Ministral-3-14B-Instruct-2512-mlx-mixed_4_8