🇺🇸 USA · Midjourney
Status: 🟩 COMPLETE 🟦 LIVING Last updated: 2026-06-26 Plain-English tagline: Independent AI image-gen lab with the most-distinctive artistic aesthetic. Originally Discord-based; now also web app. Beloved by artists / designers for its style; less faithful to prompts than Imagen / gpt-image-1.
Front-matter facts
| Field | Value |
|---|---|
| Vendor | Midjourney Inc (San Francisco, USA) — independent, no major investor |
| Country / origin | 🇺🇸 USA |
| Recommended for Australian users? | ✅ Yes — fully accessible from AUS |
| Privacy summary | Free / Pro tier images are PUBLIC in Midjourney community; private mode requires Pro tier+ |
| Free tier | None — Midjourney is paid-only |
| Paid tiers | Basic US30/mo · Pro US120/mo · Yearly discount ~20% |
| First released | July 2022 (beta) |
| Last reviewed | 2026-06-26 |
| Official site | https://midjourney.com |
What it is
Midjourney is an independent AI image generation lab and product — famous for its distinctive artistic aesthetic. Founded 2021 by David Holz (former Leap Motion co-founder), Midjourney is genuinely independent — no major venture funding, profitable, ~40-person team.
Distinguishing characteristics:
- Most-distinctive aesthetic — Midjourney outputs are often recognisable as Midjourney; the model has artistic “taste”
- Originally Discord-based — entire interface was Discord bot commands (
/imagine prompt) — unusual but became a community feature - Web app since 2024 — alphas.midjourney.com / midjourney.com — finally allows non-Discord workflow
- V6 / V6.1 / V7 — model generations, each marked quality leaps
- Style references —
--srefto match a reference image’s style - Character references —
--creffor consistent characters
Midjourney’s positioning is artistic rather than utilitarian. Where Imagen and gpt-image-1 prioritise prompt-faithfulness, Midjourney prioritises aesthetic beauty (sometimes at the expense of literal accuracy to prompts).
What you’d use it for
- Concept art / mood boards — Midjourney shines
- Stylised illustrations for any creative project
- Marketing visuals with distinctive look
- Album covers, posters, book covers
- Hobby creative work
- Pre-visualisation before live photo shoots
- Brainstorming visuals with curated taste
When NOT to use Midjourney:
- For exact prompt-faithfulness (Imagen / gpt-image-1 better)
- For text in images (Imagen / gpt-image-1 better)
- For commercial-safety (Adobe Firefly indemnifies)
- For free tier (no Midjourney free)
- For batch / API use (Midjourney is interactive, less API-friendly than others)
How to use from Australia
- Sign up at midjourney.com — subscription required
- Two interfaces:
- Web app — alphas.midjourney.com (recommended; cleaner UX)
- Discord bot — original interface; still works, useful for community
- Try a prompt:
a kookaburra in cyberpunk style, neon, rain, cinematic - AUS card accepted; Stripe-based subscription (USD with conversion)
What it costs
Basic — US$10/month
- ~200 image generations/month (3.3 GPU-hours)
- Public images
- General Commercial Terms
Standard — US$30/month
- 15 GPU-hours/month (~900 images)
- Unlimited “relax mode” generations
- Public images
- Most common tier
Pro — US$60/month
- 30 GPU-hours
- Stealth mode (private images)
- Power users / professionals
Mega — US$120/month
- 60 GPU-hours
- Top-tier creators
Annual discount
- ~20% off all tiers
- AUD via card with USD conversion
How it compares to alternatives
| Capability | Midjourney | Imagen 4 (Google) | OpenAI Image (gpt-image-1) | Adobe Firefly | Flux (BFL) |
|---|---|---|---|---|---|
| Distinctive aesthetic | Best (Midjourney-style) | Photorealistic | Versatile | Versatile | Photoreal |
| Photorealism | Good | Best | Excellent | Excellent | Excellent |
| Instruction-following | Loose interpretation | Excellent | Best | Strong | Strong |
| Text in images | Improving (v7) | Excellent | Best | Good | Good |
| Style references | Best (sref / cref) | Good | Strong | Good | Good |
| Free tier | None | Yes (via Gemini) | Yes (limited via ChatGPT) | Yes (limited credits) | Limited (via fal / Replicate) |
| Commercial safety | Standard | Strong | Strong | Best (Firefly indemnifies) | Open data |
| In-chat integration | No (Discord / web) | Yes (Gemini) | Yes (ChatGPT) | Inside CC apps | No |
| Price (consumer) | US$10-120 | Free / bundled | US$20+ (ChatGPT Plus) | US$9-30 | Per-call |
| Best for | Artistic / stylised work | Photorealism + Workspace | Chat-integrated + text-in-images | CC users + commercial safety | Open-source / customisation |
For distinctive artistic aesthetic, Midjourney is unmatched among the major options.
Privacy / data handling
- Free / Basic / Standard images are PUBLIC — visible in Midjourney community / on Explore page
- Stealth Mode (Pro+) makes images private
- Don’t generate confidential / sensitive imagery on non-Stealth tiers
- Commercial terms: outputs generally usable commercially (verify current ToS)
Recent changes
- 2026: V7 model; web app matured
- 2024: Web app launched (graduating from Discord-only)
- 2023: V5 / V6 quality leaps
- 2022: Initial beta launch
Gotchas
- Discord-only history — many tutorials reference Discord commands; the web app is newer
- Paid-only — no free tier (vs Imagen / gpt-image-1 / Firefly which all have free access)
- Public default is genuinely public — verify Stealth Mode setting if doing private work
- For exact compliance with prompts, Midjourney is the least-faithful of the major image gens
- Style is opinionated — Midjourney outputs are recognisable; for “neutral” generation, Imagen / Firefly better
- API access — Midjourney historically resisted API; third-party wrappers exist but unofficial
- AUS billing: USD via card; foreign-transaction fees on most AUS cards
See also
- Imagen (Google) 🟩 🟦
- OpenAI Image (gpt-image-1) 🟩 🟦
- Adobe Firefly 🟩 🟦
- Stable Diffusion 🟥
- Flux (Black Forest Labs) 🟥
- Ideogram 🟥
- Recraft 🟥
- Canva + Magic Studio 🟩 🟦
- Adobe Photoshop AI 🟩 🟦
- Microsoft Designer 🟩 🟦
- Multimodal (vision, audio) 🟩 🟦