Midjourney V7 Cheatsheet
A practical, behavior‑first guide to image & video generation
Last updated: February 2026 – always double‑check Midjourney’s official docs for the latest details.
Getting started with Midjourney
Important: Midjourney doesn’t offer a permanent free tier. You’ll normally need a paid subscription to generate images or videos, though occasional promos or student discounts may apply, so always check the current pricing page and student offers if relevant.
Your first week strategy
If your company’s annual gross revenue exceeds about $1M, Midjourney’s current terms require a Pro or Mega subscription. Always confirm this against the latest plan comparison and the official Terms of Service.
Common enterprise use cases
Marketing & advertising
The challenge: Campaigns and social assets need fresh visuals; stock looks generic fast.
Midjourney’s role: Generate on‑brand concepts in minutes, then personalize toward your house style and produce variations for testing. For inspiration, see prompt guides such as YUV.AI’s Midjourney V7 masterclass and DataCamp’s V7 guide.
Product visualization
The challenge: You need lifestyle shots before physical samples or full shoots exist.
Midjourney’s role: Place products into seasonal, contextual scenes to test routes and messaging long before you book a studio. See practical examples in the resources above and Superside’s prompt guide.
Editorial & publishing
The challenge: Every article deserves its own illustration, but custom art is slow and expensive.
Midjourney’s role: Quickly generate visual metaphors and scene‑based illustrations tuned to each story, without traditional stock licensing limits. See this Midjourney overview or the V7 guides linked above.
Concept design & R&D
The challenge: Teams need to see 50 options to choose 3, but only have time to sketch 5.
Midjourney’s role: Flood the room with plausible options, so human designers can spend their energy refining the winners instead of producing volume. See design‑oriented takes like Lily’s V7 tips.
Where you work: Discord vs Web
Discord (original method)
- Use bot commands in channels or DMs, e.g.
/imagine prompt: your description. See basics in the Midjourney entry. - Great for inspiration and community learning; generations are visible unless you use stealth options.
Web interface (easiest for most people)
- Visual gallery at midjourney.com with editor tools like pan, zoom, and vary region; see coverage in this feature explainer and Tom’s Guide on V7.
- Easier to organize projects, moodboards, and brand‑safe libraries; recent updates like moodboard homes are described on Midjourney’s update blog.
Think of Midjourney as a colleague: give it clear briefs, review quickly, and log what works. The habit of “generate → review → rate” will move quality more than memorizing every parameter. This reflects the behavior‑first approach of Conor Grennan’s AI Mindset.
Key V7 capabilities
- V7 improves hands, faces, and complex scenes versus earlier models; see technical and press coverage such as Tom’s Guide and Engadget.
- Better adherence to subtle style cues and references, discussed in practical V7 tips and prompt masterclasses.
- Much faster, cheaper generations for early exploration; see this feature update and workflow tips.
- “Enhance” lets you re‑render a draft at full quality once you pick a favorite, highlighted in multiple V7 guides.
- Rate images; after roughly 200 ratings, V7 leans toward your preferences by default. This behavior is described in Midjourney’s Version notes and community guides like Lily’s tips.
- Use controls like
--p 0when you need more neutral output for certain projects.
- Turbo: fastest, highest quality; most GPU‑intensive, as noted in V7 coverage.
- Fast: default priority for most work; see official plan description.
- Relax: slower but unlimited on Standard+ plans, per the same plan comparison.
- Draft: fast, low‑cost ideation mode before finals; see the V7 feature explainers above.
- Use reference images or SREF codes to keep campaigns visually coherent; see style reference tips and prompt masterclasses.
- Blend multiple references to find a shared “visual language” for a product line.
-
Use
--stylize(or--s) to nudge results from literal to more artistic; moderate ranges (for example around 50–100) are a good starting point, as described in community guides like Lily’s. -
Combine with
style rawwhen you need maximum prompt adherence over default aesthetics; see detailed prompt guidance in V7 tips and DataCamp’s tutorial.
Generate 10–20 Draft variations, review as a team, then only “spend” Fast or Turbo time on the 1–3 winners. Treat GPU the way you treat studio time: burn it only once the concept is clear. This workflow is echoed in many V7 best‑practice guides.
V7 is the current default model and handles most day‑to‑day work. Some functions (like certain upscalers and zoom behaviors) still rely on V6.1 components under the hood, so results can occasionally feel slightly different between tools. Always refer to Midjourney’s Version page for the latest feature‑by‑feature breakdown.
Video generation
Context: Midjourney’s V1 Video model, launched in mid‑2025, turns a single image into short, animated clips using “Animate” controls in the web UI. See coverage like PetaPixel, Outlook Business, and Talent500.
How it works (conceptually)
1. Generate or upload an image.
2. Click Animate in the web interface.
3. Choose motion style (Automatic or describe specific movement, such as “camera slowly pans right over the city”), as shown in early V1 demos and explainers.
4. Choose motion intensity (subtle “ambient” motion vs more dynamic subject + camera movement), emphasized in video model reviews.
5. Midjourney renders short video clips; the original V1 launch supported around 5‑second clips with options to extend the sequence duration. Exact limits may change, so check the latest video docs or articles like those above for current lengths and batch behavior.
Video jobs consume significantly more GPU than stills, so behave like a producer: lock the image look first, then animate only approved frames. That habit alone saves most teams from burning through Fast hours, as highlighted in practical reviews such as Talent500’s and other 2025–26 guides.
Plan and Relax rules for video
As of early 2026, unlimited video generation in Relax Mode is only available on Pro and Mega; Basic and Standard use Fast time for video. Always confirm this against the current “Comparing Plans” page before committing a team to a tier.
Use cases
Social posts that stop the scroll
Start with a strong still, then animate only gentle camera or environmental motion. Design it to work with sound off—most feeds autoplay silently, as many social‑focused AI video guides point out.
Pitch & concept reels
Turn key frames from decks into short, animated beats. You’re not making final TVCs; you’re making story moments that help stakeholders “feel” the idea. This mirrors advice in V1 first‑look reviews.
Midjourney videos don’t ship with audio, so think like a designer: build loops that make sense muted, and add music or voiceover later in your video editor of choice. This is emphasized in articles like Talent500’s V1 overview.
Practical workflows
Campaign asset sprints
Product visualization pipeline
Editorial illustration in hours, not weeks
If you treat rating images like brushing your teeth—small, daily, automatic—you unlock personalization quickly and keep your “AI colleague” learning your taste. Tag and store strong results to seed future work; this habit is encouraged in community V7 tips and aligns with AI Mindset’s focus on behavior change.
Plans & what they’re good for
Numbers change over time, but this gives you a ballpark of how the four main plans compare. Always check Midjourney’s current pricing and plan comparison and up‑to‑date breakdowns such as GlobalGPT’s 2026 overview.
| Plan | Monthly | Approx. annual | Fast GPU hours | Image Relax | Video Relax | Best for |
|---|---|---|---|---|---|---|
| Basic | ≈$10/mo | ≈$96/year (≈$8/mo) | ~3.3 hours (≈200 images, depending on mode) | ❌ No | ❌ No | Testing, occasional personal use |
| Standard | ≈$30/mo | ≈$288/year (≈$24/mo) | ~15 hours Fast | ✅ Yes (unlimited) | ❌ No | Regular creators, content teams focusing on images |
| Pro | ≈$60/mo | ≈$576/year (≈$48/mo) | ~30 hours Fast | ✅ Yes (unlimited) | ✅ Yes (unlimited video Relax) | Agencies, freelancers with steady client work, video experimentation |
| Mega | ≈$120/mo | ≈$1152/year (≈$96/mo) | ~60 hours Fast | ✅ Yes (unlimited) | ✅ Yes (unlimited video Relax) | High‑volume teams and studios, heavy video use |
Pick a plan based on behaviors, not aspirations: estimate weekly Fast use, decide how often you’ll really need video, and start with the smallest plan that comfortably covers your current projects. You can always scale up once the habit sticks. This mirrors recommendations in 2026 plan reviews and AI Mindset’s focus on real‑world behavior.
Prompting in V7
V7 rewards natural, specific language more than clever keyword lists. Think of prompts as briefing a collaborator, not programming a machine. See detailed examples in YUV.AI’s V7 course, Lily’s 20 tips, and Superside’s prompt guide.
A simple structure that works
- Subject: Who or what are we looking at?
- Scene: Where are they? What’s the environment?
- Styling: Lighting, mood, lens, era, medium.
- Intent: Why are you generating this? (Ad hero, thumbnail, editorial opener.)
Example: “Portrait of a thoughtful founder in a warm, modern office, soft window light, 35mm lens look, subtle depth of field, editorial style for a magazine cover story.” This style of prompt is consistent with the examples in the guides linked above.
In an AI Mindset frame, the win is changing your team’s behavior: write clearer briefs, reflect on what worked, and iterate based on outcomes. The tool gets better as your habits do. See AI Mindset and talks such as Conor Grennan’s conversations on practical AI for this philosophy in action.