Qwen3.5-122B-A10B optimized for MLX.
This quant supports image input and requires a vision-enabled MLX server.
1# Start server at http://localhost:8080/chat/completions
2uvx --from mlx-vlm --with torchvision \
3 mlx_vlm.server \
4 --host 127.0.0.1 \
5 --port 8080 \
6 --model spicyneuron/Qwen3.5-122B-A10B-MLX-4.7bit-vision
Quantized using a custom script inspired by Unsloth/AesSedai/ubergarm style mixed-precision GGUFs.
MLX quantization options differ than llama.cpp, but the principles are the same:
