Reconsider Image Generation Classification

I think it is not necessary to separate Image Edit and Image Generation.
Image Generation models are actually Text2Image models.
Image Edit models are actually TextImage2Image models.
Recent advanced models are TextImage2Image models with image reference optional.
It is a trend to mix the modality of image and text as input while image is optional.
In the api the only difference is whether to accept/require image (file or base64 string)
So maybe we can put all image generation models together and tag them with the input modality right? (Same as the T&I2V and  T2V models)
As venice is trying to add more advanced proprietary anonymized models, it's really important for the team to reconsider it.