Model
Midjourney
Run Midjourney v7, v8 and Niji with style reference, omni reference and image-to-video. Four images per generation.
Midjourney remains the model with the strongest aesthetic default — output looks composed rather than merely correct, which is why it is still the first choice for anything where the image has to be attractive rather than accurate.
v7 and v8 are available alongside Niji, the anime-tuned variant. The reference endpoints matter most in practice: style reference carries a look across generations, and omni reference carries a subject, which together are how you get a consistent series rather than isolated images.
Midjourney returns four images per job, all stored separately in the gallery so you can keep one and discard the rest. It is premium-tier at 1.5x.
FAQ
Midjourney — questions
Which Midjourney versions are available?
v7 and v8, plus Niji for anime styling, and v7 style-reference, omni-reference and image-to-video endpoints.
Do I get four images like in Discord?
Yes — a generation returns four images, each stored as its own artifact in your gallery.
How do I keep a consistent style across images?
Use style reference to carry a look, or omni reference to carry a subject, across multiple generations.
Keep reading
Related guides
Start generating today
100 free credits on signup — roughly fifty images, or twenty-five voiceovers, or three premium video clips.
No card required.