Keep the "old" Stable Diffusion đ The quality is not comparable to the VeniceSD
As I stated in an ealier request - I think itâs goo to keep the old SD, as an artist Iâm really into creating special almost magical art. The prompts just donât seem to be pickedup well. Below the new SD and beneath the new. Itâs just to ârealisticâ and not according the prompt (plus parts are just missing). On top of that I got a âwarmingâ that certain words (like girl) are not allowed so have to use âladyâ, so weird and not free, nothing explicit on this image IMO.

Log in to comment and vote
Comments2
Sapphire Mirror
Jul 17, 2025
Yep, I got my subscription today because being uncensored was the selling point over ChatGPT or Grok, which meant that I wouldnât have to worry about things being âtoo spicyâ when making something, and after the AI telling me that I couldnât get any image of people whatsoever (even fantasy non-humans) I canceled. The AI says if done in 24 hours I can get a refund, but I wonât hold my breath. Given how others already had accounts they probably donât have other AI models they are using and are now having to scramble to transfer things over, I feel bad for them. Every AI service feels the heat when they start getting big, and this response tells me they are not cut out to handle it in the slightest.
Jade Star
Jun 15, 2025
(the sroty continues) The Venice image capabilities are getting worse and worse how I experience it lately. Look at the image below and in particular the prompt. Why does this get censored? Explain me.Â
Besides that phasing out the real SD stresses me out - as pointed out ealier the VeniceSD doesnât come close to the quality, magic and style of the real one. Keep it please.
Also the new way of upscaling is not very efficient from a UX point of view. The new upscale uses a new prompt / action for which you have to wait. Before they went in the background as proces and I could continue generating variations. Bad UX plus it takes more clicks, which is not healthy from a usability / accessibility point of view, as it requires way more clicks.
Please consider this.