How we review: this assessment is based on the product’s official documentation, current pricing and published independent evaluations; it is not a hands-on test. Read our review methodology.
Midjourney has been the reference for AI image aesthetics since its early versions, and in 2026 it shipped a new generation. V8.1 replaced V7 as the default on 11 June, and V8.2 followed as the default on 24 July, according to Midjourney’s updates page and documentation. This review assesses the product as Midjourney documents it today.
What it is
Midjourney is an independent research lab’s image generation service. You describe an image in text — optionally adding reference images, a style or personalization — and it produces a grid of candidates you can vary, upscale, edit or animate. It runs in a web app and on Discord. Our Midjourney directory entry has a quick overview.
Key features
- V8 models. Midjourney describes V8.1 as smarter and more coherent than V7, with better adherence to detailed prompts and improved text rendering. Its documentation says V8.2, the current default, focuses on aesthetics, image quality and personalization. Older versions remain selectable.
- HD mode. With HD on, Midjourney says V8.1 produces images at twice the size and four times the resolution of V7. It quotes about 4 seconds per image in standard mode and about 12 seconds in HD.
- Edit Model. Introduced with V8.2, it replaces the earlier Omni Reference, Character Reference and Retexture tools with one editing model.
- Style consistency. Style references, personalization profiles and aesthetic settings carry over between V7 and the V8 line, according to the V8.1 announcement.
- Video. Image-to-video creates 5-second clips from a starting image, with Low Motion as the default and High Motion as an option; generation defaults to four videos per prompt, which can be reduced to save GPU time.
- Fast, Relax and Turbo modes. GPU speed modes trade speed for subscription time; Midjourney’s docs note Turbo was not supported on V8.1.
- Stealth mode. Hides creations from other users on the website, on Pro and Mega only. It does not hide images generated in shared Discord channels.
Pricing
At the time of writing (October 2026), Midjourney’s plan documentation lists four subscriptions:
| Plan | Monthly | Annual (per month) | Fast GPU time | Relax mode | Stealth mode |
|---|---|---|---|---|---|
| Basic | $10 | $8 | ~3.3 hours | No | No |
| Standard | $30 | $24 | 15 hours | Yes | No |
| Pro | $60 | $48 | 30 hours | Yes | Yes |
| Mega | $120 | $96 | 60 hours | Yes | Yes |
Annual plans are paid up front. Midjourney’s terms also require companies grossing more than $1,000,000 a year to be on Pro or Mega to use images commercially. Unused Fast GPU time does not roll over. Extra Fast hours can be purchased; Midjourney describes hourly prices as experimental, so we have not quoted one.
Strengths
Aesthetic quality is Midjourney’s long-standing advantage. Independent comparisons and community rankings have consistently placed it near the top for visual style and composition, and the company’s V8 release notes target its historical weaknesses — prompt adherence and text — rather than only polish. Our guide to choosing an AI image generator explains how these strengths compare with tools that focus on control or licensing.
Style continuity is valuable for working creatives. The fact that personalization profiles and style references carry over from V7 means artists who have tuned their workflow do not have to start again with each model.
Plans are simple. Four tiers, a clear GPU-time model and a 20% annual discount make Midjourney easier to budget than many credit-based competitors, and video is included in the same subscription rather than sold separately.
Limitations
There is no free plan, so trying Midjourney means paying at least $10. Basic also lacks Relax mode, so once its roughly 3.3 Fast hours are gone, generation stops until renewal or a top-up.
Privacy costs extra. Images are visible on the website by default, and Stealth mode is reserved for the two most expensive plans; even then, images made in shared Discord channels remain visible. For client work under NDA, that is a real constraint.
Control and licensing are weaker than enterprise-oriented rivals. Midjourney does not market brand-safe training data or IP indemnity the way some design platforms do, and our explainer on how diffusion models generate images notes the copyright debates around training data that buyers should weigh.
Feature support also varies by version — Turbo mode, for instance, was not available on V8.1 — so users should check the docs before relying on a parameter.
Who it’s for
- Illustrators, concept artists and designers who prioritize visual quality and style.
- Marketers and content creators producing mood boards, campaign concepts and social imagery.
- Hobbyists willing to pay for the most distinctive results.
It is a weaker fit for teams needing private-by-default generation on a small budget, or strict commercial indemnity.
Alternatives
- Adobe Firefly — integrated with Creative Cloud and positioned for commercial safety.
- Canva — image generation inside a full design tool for non-designers.
- FLUX.1 [dev] — an open-weight model you can run yourself.
Verdict
Midjourney earns 4 out of 5. Its documented V8 improvements, HD output, unified Edit Model and bundled video keep it the leading choice for aesthetic image generation, and its four-tier pricing is easy to understand. We deduct a point for the lack of a free tier, public-by-default images with Stealth limited to the top plans, and weaker licensing assurances than enterprise-focused rivals.
