They also just released v5.1, which seems to be quite a bit better than v5.

Is there a comparison? It is interesting that there do not seem to be any Midjourney benchmarks. E.g.

https://paperswithcode.com/sota/text-to-image-generation-on-...

Parti and Imagen are still on top, followed by Dall-E 2.

If their model is so great, why are they afraid of benchmarks?

I literally just don't feel like running them tbh, and see no reason to publish them either way. Mostly prefer to let the outputs speak for themselves.

For a while I was using an FID variant for evaluation during training, but didn't find it very helpful vs just looking at output images.

Okay. That's probably the difference between a commercial and a research project.

Agreed. I pay for MJ and have several SD versions running on my PC. I like the ability to fine tune the SD models and my Pc with a 4090 is plenty fast, but I can't match MJ's output on artistic quality. SD allows for 4k sized outputs which is great but I can't use the art like I would like. FWIW the SD NSFW community is large but that is not where I invest my time with AI art.

Generate the following picture in mid journey: "A school of dolphins spanking a mermaid with their flukes."

A 1000 V-rolls won't get you there. For something like this control net combined with inpainting is indispensable. Not to mention the excessively heavy handed censorship in MJ.

Midjourney excels in overall quality, but it completely falls down if you have an actual complex vision.

It seems Midjourney is great at generating non-pornographic pictures.

Uhh... Spanking isn't inherently pornographic - that particular image was supposed to be a Gary Larson parody style comic.

Here have another prompt: "Rapunzel has let her hair all the way down a tower. The hero has been tied up by the witch, and annoyed at Rapunzel's continual attempts to escape, the witch throws the bottom of her hair into a paper shredder at the base of the tower."

You could v-roll until the heat death of the universe without even getting close.

Midjourney is great if all you're capable of conceiving is 90s Mad Magazine templatized mad Libs, banal crap like "Darth Vader as a French pantomime street artist".

Unfortunately that also describes the majority of midjourney users.

Yeah, prompts which basically just list properties (Darth Vader, French, pantomime, street artist) seem to work well, but relations are mostly too hard for these models. Even "a monk playing chess against a clown" or "a blue book on top of a yellow book" is out of reach for Bing's Dall-E ~3, and Midjourney probably isn't much better here.

https://www.bing.com/images/create/a-monk-playing-chess-agai...

https://www.bing.com/images/create/a-blue-book-on-top-of-a-y...

Simple (prompt only) use of generative models is quite good at creating simple artistic pictures you might actually hang on a wall. Trying to create a complex scene with just a prompt seems still a few years off though.