This Midjourney review examines one of the best-known AI image generation platforms in 2026, including its current V8.2 model, editing tools, style controls, video features, pricing, privacy, commercial-use rules, and major alternatives.
Midjourney has changed considerably from the Discord-centered image generator many people still associate with the name. Today, users can create directly on the web, edit generated or uploaded images, develop reusable visual styles, organize projects, and turn still images into short videos.
Its biggest appeal remains visual creation. Midjourney is designed for people who want to explore concepts, compositions, moods, illustrations, photography-inspired imagery, and other creative directions quickly.
But choosing it in 2026 involves more than deciding whether you like its images. GPU limits, public-by-default creations, Stealth access, commercial-use terms, editing limitations, and the lack of a general public API can all affect whether it fits your workflow.
Review basis: this is a research-based review of documented features and terms, rather than a hands-on V8.2 benchmark. Practical recommendations reflect workflow requirements, not independently measured image-quality rankings.
What Is Midjourney?
Midjourney is a paid generative visual creation service operated by Midjourney, Inc. It can generate images from text prompts and visual references, edit or extend images, explore reusable visual styles, and animate a still image into a short video.
The platform is relevant to artists, illustrators, concept designers, marketers, social creators, agencies, and other users who need rapid visual exploration.
There are two main ways to use it:
- Midjourney website: creation, editing, style exploration, personalization, Moodboards, organization, and downloads.
- Discord: the Midjourney Bot, slash commands, shared generation, community interaction, and familiar command-based workflows.
Discord is no longer required to start using Midjourney. New users can sign up with Google and work directly on the website. Discord is only necessary if you specifically want the Discord workflow.
This is important because many older tutorials still describe Discord as mandatory or suggest that users must generate a certain number of images before gaining web access. That information is outdated.
Midjourney V8.2 and Image Generation
V8.2 is the current default general Midjourney model, becoming the default on July 24, 2026, according to the official model-version documentation. Earlier versions remain relevant to some workflows, while Niji 7 provides a separate path focused on anime and Eastern illustration styles.
A normal image request produces four alternatives. On the website, these appear as individual images rather than the traditional Discord grid.
V8.2 standard-definition square images begin at 1024×1024 pixels. HD generation can produce 2048×2048 square output, with dimensions changing according to the selected aspect ratio.
Existing SD images can also use Subtle or Creative upscaling. Creative Upscale may alter details rather than simply enlarging pixels, so it should be inspected when exact details matter.
Midjourney offers several controls for steering a generation. These include aspect ratio, Stylize, Raw mode, variety or chaos, Weird, negative instructions through --no, and seeds where supported.
Raw mode and lower Stylize values can be particularly relevant when you want the system to follow a prompt more literally instead of applying as much of its default aesthetic direction.
Not every parameter works with every model. Older prompt guides can therefore be misleading when they mix V7, V8, and legacy controls.
Draft Mode
Draft Mode offers a lower-cost way to explore ideas before committing to a finished image. The current dedicated help page describes V8.1 and V8.2 web drafts as batches of 24 lower-resolution images, using 0.4 GPU minutes per prompt. Selected drafts can then be regenerated at SD or HD resolution.
This workflow is useful for comparing compositions quickly, but it differs from V7 Draft Mode and is not currently available for V8.1 or V8.2 in Discord. Draft outputs should be treated as starting points rather than final production assets.
Style References, Moodboards and Personalization
One of Midjourney’s more distinctive areas is visual style control.
Style Reference allows an image or style code to influence the visual treatment of new generations. It is intended to guide aesthetics rather than copy the identity or exact composition of the reference.
Moodboards let users build broader visual directions from curated sets of images. They can be useful when developing a campaign mood, illustration direction, story world, or repeated visual language.
Personalization uses preferences informed by image ratings to influence how the platform interprets your aesthetic taste.
Together, these tools make the service particularly interesting for iterative art direction. Instead of attempting to describe every aesthetic characteristic repeatedly in text, users can combine prompts with visual references and personalized style direction.
However, stylistic consistency should not be confused with exact reproducibility. A reference can influence an output without guaranteeing an identical character, product, logo, face, layout, or piece of typography across multiple generations.
Image Editing and Subject Consistency
Midjourney is no longer limited to generating a new image from scratch.
The web Editor can work with generated or uploaded images and supports selected-area editing, erasing, moving, resizing, layers, inpainting, and outpainting.
The newer V8 Edit Model also supports natural-language editing instructions and can accept up to four reference images. This can help when trying to preserve a character, object, or other visual element while changing the surrounding image.
Reference-based subject preservation is still generative rather than a guarantee of identical details. In selected-area Editor workflows, visible regions outside the regenerated areas can remain unchanged, but broad natural-language edits and subject references should not be treated as exact identity locks.
If exact product packaging, a person’s appearance, a logo, typography, or another identity-critical element must remain unchanged, every output should be reviewed manually.
There are also model-specific compatibility restrictions. For example, current V8 Edit Model results are not compatible with Remix even though ordinary V8.2 generations can support Remix.
Resolution depends on the editing route. The dedicated Edit Model documentation lists support for --hd, while guidance for some inpainting and outpainting tools describes SD results that can be upscaled afterward. Check the dimensions of the downloaded result rather than assuming every edit preserves the source resolution.
How Well Does Midjourney Handle Text?
Midjourney can generate text inside images, but exact typography remains something to proofread rather than trust automatically.
The company’s own guidance recommends putting short required text inside double quotation marks and using approaches such as Raw mode or lower Stylize when exact wording is difficult.
This makes generated text potentially useful for concepts, mockups, posters, or visual experimentation, but final advertising headlines, prices, legal text, brand names, and logos should be checked carefully.
For production design, it can be safer to generate the visual composition first and add mission-critical typography later in a dedicated design application.
Midjourney Image-to-Video
Midjourney now includes image-to-video generation, but it should not be confused with a complete AI filmmaking platform.
The current workflow begins with one generated image or an uploaded starting frame and turns it into a five-second video. Users can provide an optional motion prompt, choose lower or higher motion, create loops, and work with end-frame options.
A clip can be extended by four seconds up to four times, resulting in a maximum documented duration of 21 seconds.
All four plans support SD video in Fast mode. Standard, Pro, and Mega also support HD video in Fast mode, while SD video in Relax mode is limited to Pro and Mega. The documentation classifies SD as 480p and HD as 720p, with actual pixel dimensions varying by aspect ratio.
Video generation can consume considerably more Fast GPU time than still images, especially when requesting multiple HD videos. The default batch contains four clips; choosing a batch of one or two can reduce the GPU cost. Extensions also consume GPU time.
There is no verified native audio-generation system or conventional timeline editor in the current documentation. Creators building longer narrated videos will therefore need additional tools for editing, audio, sequencing, and production.
Midjourney Web vs Discord
For most new users, the website is now the more straightforward starting point.
The web interface provides Create, Edit, Explore, Personalize, Moodboards, and Organize without requiring users to learn Discord slash commands.
It also provides visual editing and organization features that fit naturally into an image workflow.
Discord remains useful for people who enjoy community generation, chat, slash commands, and the bot-based workflow. Existing Discord users should also be careful to sign in correctly so they retain access to their previous generation history.
The important point is that neither interface should be described as the only “real” Midjourney experience in 2026.
Midjourney Pricing in 2026
Midjourney currently offers four paid subscription tiers. The official plan comparison lists current prices, allowances, and feature access.
| Plan | Monthly | Annual | Fast GPU Time | Key Difference |
|---|---|---|---|---|
| Basic | $10 | $96 ($8/mo effective) | 3.3 hours | Entry tier, no Relax or Stealth |
| Standard | $30 | $288 ($24/mo effective) | 15 hours | Unlimited image Relax |
| Pro | $60 | $576 ($48/mo effective) | 30 hours | Stealth and SD video Relax |
| Mega | $120 | $1,152 ($96/mo effective) | 60 hours | Higher Fast capacity and video concurrency |
*Public U.S. pricing verified in October 2026. Annual subscriptions are paid upfront. Taxes and checkout terms can vary.
All plans can purchase additional Fast GPU time at $4 per hour. Included Fast time resets each monthly subscription cycle, including on annual billing, and unused included time does not roll over. Subscriptions renew automatically unless canceled.
Fast vs Relax
Fast mode spends the GPU time included with your subscription.
Relax uses a slower shared queue instead of consuming normal Fast minutes for eligible work. Standard introduces unlimited Relax for still images, while Pro and Mega extend Relax access to SD video.
“Unlimited” therefore does not mean unlimited fast generation. Relax jobs can wait in a queue and remain subject to service limits.
GPU hours should also not be interpreted as a guaranteed number of images. Different operations consume different amounts of compute.
Is Midjourney Free?
No ongoing free Midjourney plan is currently available on the website or through Discord.
The official documentation also says there is no general free web or Discord trial.
A limited trial is documented for the separate niji · journey mobile application on iOS and Android, but that should not be presented as a general Midjourney free tier.
For someone simply wanting to test the main product, Basic at $10 per month is therefore the normal entry point.
Which Midjourney Plan Makes the Most Sense?
Basic is suitable for occasional users who want to experiment with the platform and can work within 3.3 hours of monthly Fast GPU time.
Standard is more attractive for regular image creators because unlimited image Relax changes the economics of experimentation. You can reserve Fast time for work that needs speed and use Relax for less urgent generations.
Pro becomes important when privacy matters because it is the first plan with Stealth. It also adds more Fast GPU time, higher concurrency, and Relax access for SD video.
Mega primarily makes sense for substantially heavier workloads requiring more Fast time and concurrency.
Buying a higher tier does not inherently make an individual image better. The major differences are capacity, workflow access, privacy, and concurrency.
Commercial Use, Ownership and Stealth
Midjourney’s current Terms of Service say users own the assets they create to the fullest extent possible under applicable law, subject to the contract and third-party rights. Upscaling someone else’s image does not transfer ownership of that image to you.
Commercial use is generally available to paid subscribers, but there is an important business condition: a company—or an employee creating on behalf of a company—with more than $1 million in annual gross revenue must use Pro or Mega to own assets under the applicable terms.
Contractual ownership should not be confused with guaranteed copyright protection. Copyrightability of AI-generated material can depend on jurisdiction and the level of human authorship.
Users are also responsible for rights associated with their inputs and outputs. Ownership of eligible assets continues after canceling or downgrading the subscription.
Midjourney also receives a broad license: the terms grant it ongoing, non-exclusive rights to use uploaded content and generated assets, including reproduction, distribution, and sublicensing. That license survives the end of the agreement. Stealth controls public visibility; it should not be read as removing this contractual license.
Creations Are Public by Default
This is one of the most important considerations for professional users.
Midjourney creations may be publicly discoverable and remixable on its website by default—even when a job was created through a Discord direct message or private server. Current V8 Edit Model creations, including edits based on external images, also appear in the website gallery unless Stealth is enabled.
Stealth Mode, available on Pro and Mega, can prevent eligible jobs created while it is enabled from appearing publicly on the Midjourney website.
But Stealth is not the same as complete confidentiality. It does not hide something you deliberately post in a public Discord channel, does not automatically hide every older creation, and does not prevent Midjourney itself from processing content under its terms.
For confidential client work, unpublished products, sensitive campaign material, or proprietary visual references, review visibility and contract terms before uploading anything—not after generating the final image.
Midjourney Pros and Cons
Pros
- Strong visual exploration: useful for rapidly developing different moods, compositions, and creative directions.
- Rich style controls: Style References, Moodboards, Personalization, and prompt controls support repeatable art direction.
- Full web workflow: Discord is no longer required for normal onboarding and creation.
- Growing editing capabilities: uploaded-image editing, inpainting, outpainting, layers, and reference-based V8 edits expand the workflow beyond prompting.
- Useful iteration tools: variations, Remix, upscaling, folders, and Draft workflows help explore ideas efficiently.
- Image-to-video: still concepts can be turned into short motion explorations without leaving the platform.
Cons
- No normal free plan: the main web and Discord product requires a paid subscription.
- Public by default: privacy-sensitive work may require the more expensive Pro or Mega plan and careful workflow management.
- Exact consistency is not guaranteed: references influence generations rather than locking every subject or detail.
- Text still needs checking: important typography should be proofread or finished in another design tool.
- GPU pricing takes time to understand: Fast, Relax, resolution, video, and different operations affect actual usage.
- No general official public API: it is a poor fit for developers who need supported automated image generation inside production applications.
Who Is Midjourney Best For?
Midjourney is particularly well suited to creators who value rapid visual ideation and art direction.
Concept artists and illustrators can explore scenes and styles. Designers and agencies can create mood directions and campaign concepts. Marketers and social creators can generate visual drafts in multiple aspect ratios, while storyboarders can use stills and short motion clips to explore ideas.
It becomes less suitable when the main requirement is exact production control.
Locked packaging, precise brand systems, final vector logos, exact typography, long-form video editing, confidential client material, or automated developer integration may require additional or alternative tools.
Midjourney Alternatives
The right alternative depends on what you are trying to produce rather than which generator has the longest feature list.
ChatGPT Images may be more attractive when conversational iteration and natural-language editing are central to the workflow.
Adobe Firefly and Photoshop are worth considering when generative imagery needs to flow into a more traditional layer-based photo and design workflow.
Ideogram is relevant for users focused heavily on text-forward graphics, posters, and editable design layouts.
Recraft deserves consideration when native vector and SVG workflows, icons, or production-oriented design assets are important.
These are workflow distinctions rather than a universal ranking. A creator prioritizing cinematic visual exploration can have very different requirements from a designer producing editable vectors or an application developer who needs an official API.
Final Verdict
Midjourney in 2026 is a much more complete visual creation platform than its old reputation as a Discord-only image generator suggests.
V8.2, the web interface, Style References, Moodboards, Personalization, the Editor, and image-to-video tools create a workflow that extends from initial visual exploration through iteration and short-form motion.
Its strongest use case is still creative direction: quickly exploring what a scene, campaign, illustration, character world, or visual concept could look like.
Its limitations become more important as work moves toward final production. Exact typography and subject consistency require inspection, image editing remains generative rather than deterministic, video is short-form, and there is no generally available official API for production automation.
Privacy also deserves particular attention. Because creations are public by default and Stealth begins with Pro, professional users should decide how sensitive their inputs are before choosing a plan.
Basic is a reasonable entry point for occasional experimentation. Standard is more practical for regular image generation because of Relax mode. Pro becomes the important tier when Stealth, greater capacity, or SD video Relax matters. Mega is mainly a higher-capacity option for heavier use.
For users who primarily want fast, flexible visual ideation and strong style direction, Midjourney offers a deep creative workflow. For exact production design, confidential work, long-form video, or API-driven applications, its strengths are better treated as one part of a larger toolchain rather than the entire solution.
Last reviewed: October 6, 2026. Midjourney changes frequently. Model availability, GPU allowances, pricing, editing behavior, video access, Stealth rules, and commercial terms should be verified before purchasing or relying on a specific workflow.
