Pic related is what it generates when i ask for a "Dog"

I was setting SD 3.5 Large up, to generate image in my OpenWebUI. But it even generate these broken images if i just go directly to

Steps i took:
>installed Python 3.10
>installed Git for windows
>ran "git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui.git";
>Download the SD 3.5 Large model (sd3.5_large.safetensors) 16.5GB
>created a "clip" folder inside "models" folder
>downloaded clip_g.safetensors, clip_l.safetensors, t5xxl_fp16.safetensors
>checked hash of all files from huggingface, match
>Configure webui-user.bat
>set COMMANDLINE_ARGS=--api --listen
>run webui-user.bat
>go to http://127.0.0.1:7860
And then generate gliched gibberish ???

System:
>Windows 11
>RTX 4090
>Ram 128gb
>2Tb NVMe

>Nvidia Driver Version: 572.16
>Cuda version 12.8