Midjourney v7 launched in mid-2025 and matured through 2026 with major upgrades to style consistency, character consistency, and personalization. The core skill of writing a prompt didn't change, but the parameters changed enough that old habits produce worse results. Here is the current playbook.
What changed in 2026
- --sref (style references) and --cref (character references) are now production-grade. Generate a 20-image series with consistent style and recurring characters.
- Personalization (--p) is dramatically better. With 200+ ratings, Midjourney learns your taste; toggling --p on a prompt biases output meaningfully toward what you've upvoted.
- --chaos and --weird rebalanced — small values (5–25) now produce noticeable variation; large values produce truly weird output that occasionally hits.
Prompt structure that works
Modern Midjourney prompts work best in this order:
[subject], [setting], [composition], [style], [mood/lighting] --params
Example:
a tired barista at 6am, behind a marble espresso bar in a sunlit
Brooklyn cafe, three-quarter view, photorealistic film photography,
warm morning light, hint of steam --ar 3:2 --style raw --v 7
Style references (--sref)
Provide an image URL or seed; Midjourney transfers style without copying composition. Best use: produce a 10-image series with consistent visual identity.
[your prompt] --sref https://i.mj.run/XXXXX --sw 100
--sw controls strength: 0 = ignore style, 1000 = strong style. 100-300 is the sweet spot for most work.
Character references (--cref)
Same idea for characters. Generate a character once; reuse across images. Best for storyboards, comics, brand mascots.
[your prompt] --cref https://i.mj.run/CHAR --cw 100
Strength --cw: 0 = face only, 100 = strong character (face + clothes + hair).
Personalization
After 200 image ratings (Midjourney rolls these into your profile), --p biases toward what you've upvoted. Effective for users who consistently rate; less effective if your ratings are inconsistent.
--chaos and --weird
| Param |
Range |
What it does |
| --chaos |
0-100 |
Variation across the 4 grid options |
| --weird |
0-3000 |
Aesthetic deviation from "normal" |
| --stylize |
0-1000 |
How much Midjourney embellishes |
Practical defaults:
- Photorealistic:
--chaos 15 --stylize 100 --style raw
- Stylized:
--chaos 30 --stylize 500
- Exploratory:
--chaos 50 --weird 250
Style raw
--style raw reduces Midjourney's "house style" lean — closer to a photograph and to what you literally described. Default for product shots, photojournalism style, and any time the prompt is precise.
Negative prompts (--no)
Use for elements you actively don't want. --no text, watermark, logo, blur is a useful default for clean output.
Aspect ratios
--ar 3:2 for film photography, --ar 16:9 for YouTube thumbnails, --ar 9:16 for Reels/TikTok, --ar 1:1 for Instagram. v7 handles 3:2 and 4:5 better than v6 for portrait subjects.
Patterns that consistently work
Character series: generate a hero character with a strong --cref base, then vary scenarios while keeping --cref + --sref constant.
Brand asset library: lock --sref to your brand mood-board image. Generate dozens of consistent visuals.
Photorealistic product: --style raw --stylize 50 --ar 3:2. Avoid the "Midjourney glow" that screams AI on product shots.
Editorial illustration: --stylize 500 --weird 100 --chaos 25 produces the kind of stylized work editors actually buy.
Common mistakes
Over-specifying. Long prompts past 60 words start fighting themselves. Shorter, structured prompts beat long lists.
Underusing --sref. Most v6 users haven't switched to using --sref/--cref religiously. The consistency gains are dramatic.
Ignoring grid choice. Grid 1-4 of every generation differs. Picking the strongest before upscaling matters more than re-prompting.
FAQ
Can I use Midjourney commercially?
Yes on paid plans (Standard and above). Free trial output is non-commercial.
Is Midjourney still Discord-only?
No — the web app (alpha.midjourney.com) is now the primary interface in 2026. Discord still works.
What about Midjourney vs Flux Pro for 2026?
Midjourney leads on aesthetic quality and style coherence. Flux Pro leads on prompt adherence and text rendering. Many pros use both.
Where to go next
For related guides see Flux Pro prompting guide, Stable Diffusion 3.5 guide, and AI image prompt engineering in 2026.