Three Big Releases in v7
1. Draft Mode — Images now render at 1/10 cost and 5x speed. Critical for fast moodboards and concept iteration. "draft mode" toggle is now standard in the v7 prompt UI.
2. Omni Reference — Hold character, object or style consistency across different scenes. Best for ad series, lookbook shoots, and social campaign sequences. Multiple references stack (person + object + style).
3. Personalization v2 — Trains the model on 200+ images you rate to calibrate to your personal/brand aesthetic. Once set, your personal model runs behind every prompt.
Web v2 — Goodbye Discord
Midjourney ended its Discord dependency in 2025-26. The new web interface lives at midjourney.com/imagine: prompt history, folder structure, team collaboration, comment mode — all built in. Discord still works but most new features are web-only.
Pro Prompt Formulas
Ad creative: [product] photographed by [photographer name], [setting], [lighting], [mood], 50mm f/1.4, hyperdetailed, --ar 16:9 --style raw --v 7 --p [your-personal-code]
Fashion editorial: fashion editorial of [model description] wearing [outfit], shot for [magazine name], [location], [light], 85mm portrait lens, film grain, --ar 4:5 --style raw --v 7
Architectural render: architectural visualization of [building type], [environment], [time of day], [atmosphere], cinema4d + corona renderer style, hyperrealistic materials, --ar 16:9 --v 7 --q 2
Omni Reference for Character Consistency
When you need the same model across scenes in a campaign:
- 1. Upload a clean reference photo of the model
- 2. Append
--oref [URL] --ow 100(Omni Weight: 0-1000) - 3. Write the scene; the model stays consistent
- 4. Add
--sref [style_url] --sw 200for style reference
Pricing 2026 (Current)
- Basic ($10/mo): ~200 fast images, limited upscale
- Standard ($30/mo): 15h fast GPU, unlimited relax mode
- Pro ($60/mo): 30h fast, stealth mode (private gallery)
- Mega ($120/mo): 60h fast, highest priority
Annual plans get 20% discount. Draft Mode stretches fast hours 5x.
Pro Workflow — Ad Campaign
Stage 1 — Brief to Moodboard: 50-100 images in Draft mode, set direction (~15 min).
Stage 2 — Concept Polish: 10 alternative directions at standard speed (~30 min).
Stage 3 — Character Lock: Lock the final direction with Omni Reference, generate 5-7 scenes (~45 min).
Stage 4 — Upscale & Post: 4K upscale + Photoshop/Lightroom finish (~1 hour).
v7 Anti-Patterns
1. Style soup: Don't pile 5 artists/styles on one prompt. v7 produces chaos rather than fusion.
2. Parameter overload: --chaos 100 --weird 3000 --s 1000 destroys control. Tune one knob at a time.
3. Single seed dependency: Locking too tight hides variation. Let seeds vary, pick the best 3.
Midjourney v7 vs Others
vs Flux Pro: Flux wins on photoreal humans and text accuracy. Midjourney v7 leads on aesthetic consistency and mood.
vs DALL-E 3: DALL-E is practical inside ChatGPT. Midjourney v7 is the creative-quality tool.
vs ImageFX (Imagen 4): ImageFX is free and fast. Midjourney v7 is mature for commercial pipelines.
PAM Istanbul's Midjourney v7 Use
We use v7 in two main flows: (1) Pre-production moodboards — three different directions to the brand within 24 hours of brief, (2) Ad series consistency — Omni Reference to show the same character across 8-10 social scenes. Always alongside live shoot, never instead.
For an AI-supported campaign production, reach the studio here.