Art in AI – MidJourney X DALL-E 2 X StabilityAI

A cool and clear comparison. Although I find the conclusions rather misleading.

Each model has its specifics, and it's ideal to tailor the prompt accordingly. MidJourney excels with long "convergent" prompts, but there isn't a single one among those tested. DALL-E 2, on the other hand, can mimic artistic styles much better than demonstrated here, and in portrait painting, it undoubtedly leads the way for me. MidJourney is ideal for landscapes or architecture, boasting incredible detail and lighting. I also use other models, but I have no personal experience with StabilityAI.

In any case, I am amazed at how rapidly this area of AI is advancing—literally every week!

I have finally been able to test a new AI: StabilityAI. I was really looking forward to it after the creations and the comments I was seeing on social media and forums; and the truth is that it offers very interesting options. Let’s compare them!

Attached you have a comparison of MidJourney, DALL-E 2 and StabilityAI (in this last one I have used the k_lms sampler, I will explain later). I have divided it into two areas: Convergent (when we ask the AI something with a lot of detail and specifications) and Divergent (when we give it very few guidelines and let it explore). These are further divided into two:

• Convergent // Real. When what we ask it to do is something that could belong to the real world.
• Convergent // Imaginary. When what we ask for does not exist as such; they are new creations.
• Divergent // Concrete. When, although we let it explore, we are asking for a specific result.
• Divergent // Abstract. When we are asking it to give us sensations and imaginary concepts, nothing concrete.

I think the results speak for themselves, but I am ready to draw some conclusions that I find interesting:

• For imaginary and sensory outputs, MidJourney is unbeatable.
• For original, different, "creative" results, I think MidJourney is the choice.
• For realistic results (in general), I believe DALL-E 2 is the best option, although depending on the case, StabilityAI presents interesting results. I would try both. MidJourney falters because it is not capable of producing coherent creations with a lot of detail. It may appear so, but it is not.
• For human figures, if we desire a result with photographic realism (as close as possible), DALL-E 2 and StabilityAI are again the options. If we want it done in a specific artistic style and are not expecting perfection, MidJourney is once more the choice.
• I think StabilityAI occupies a middle ground between DALL-E 2 and MidJourney, being more artistic than the former and more realistic than the latter. I believe it can be a genuinely powerful option to consider. Additionally, it introduces a new variable: you can choose the method (sampler) with which you want it to generate images. For the comparison, I have always used the same, but changing the sampler depending on the creation can yield really interesting details.

We’ll have to follow them closely and keep experimenting!