Same base model and 3.6 bpw quantization — text-only primary plus the vision add-on.
Note Text primary
Note Vision add-on (3D MoonViT) on the same base