The original 3D model was generated from an AI generated image by another AI tool. But, their model was not matching the AI generated image at all. And I kind of liked the first AI generated image. A couple years later we now have very powerful AI tools that can generate super accurate and high fidelity 3D models from any image.

I used Hunyuan 2.5 to generate this 3D model + textures from the original AI generated concept image.

EDIT: I retopologized the AI output, it is now really low poly…