Black Forest Labs’ aesthetic-quality leader in AI image generation. V7 is today’s production default on web and Discord. V8.1 moved out of alpha-only on April 30, 2026 and is now available on both Discord and midjourney.com (HD is temporarily SD-by-default during a server transition; use --hd or settings to opt back in). The shift brings sharper renders, better SREFs, moodboards, and HD images. V8.2 is being trained on community rating data.
Pricing runs $10-$120/month. No free tier. No generally-available public API.
Recent developments
- April 30, 2026: V8.1 updates graduated the model out of alpha-only access. V8.1 is now reachable from both Discord and midjourney.com, with sharpness and image-quality improvements most visible on SREFs, Moodboards, and HD images. HD is temporarily defaulting to SD during a server transition, with the
--hdflag or a settings change to opt back in. Community ratings now feed V8.2 development. - April 21, 2026: OpenAI launches ChatGPT Images 2.0 and gpt-image-2 with native reasoning, 4K output, and multilingual text rendering across 12+ languages. gpt-image-2 is the closest challenger to V7’s artistic lead to date, especially on infographics and text-heavy generation. Midjourney retains the cinematic-composition crown; OpenAI bundle-economics pressure the $10/mo Basic tier’s entry-point case.
System Verdict
Pick Midjourney if visual quality is the single axis you cannot compromise on. Still the reference point for cinematic composition, lighting, and stylized illustration.
The web editor now leads Discord for most workflows. Remix, Vary Region, Pan, Zoom Out, Retexture, Smart Select, and a proper Organize page with folders, search, and bulk actions are all native. Niji 7 (launched January 2026) covers anime and illustrative styles in the same subscription. V8.1 is now reachable from both Discord and midjourney.com after the April 30 rollout.
Skip it if you need a programmatic API, a free tier, or guaranteed text rendering. The only official API is Enterprise, and it is invitation-gated. Ideogram still beats V7 on text-in-image for logos and thumbnails. The $10/mo Basic floor shuts out casual users; Leonardo and Flux are better entry points.
Who pays which tier: Basic $10/mo for hobbyists, Standard $30/mo for most working designers (unlimited Relax + 15 Fast hours), Pro $60/mo if you need Stealth mode for client confidentiality, Mega $120/mo for agencies running sustained Fast-mode workloads. Businesses grossing over $1M/year must be on Pro or Mega for commercial rights.
Key Facts
| Production default model | V7 (default since June 16, 2025) |
| Available newer models | V8 Alpha (March 17, 2026) · V8.1 graduated out of alpha-only on April 30, 2026 and now runs on both Discord and midjourney.com |
| Anime/illustration model | Niji 7 (launched January 9, 2026) |
| V8.1 output | Native 2K HD when enabled; HD is temporarily SD-by-default during a server transition, with --hd or settings to opt back in |
| In training | V8.2, fed by community rating data |
| Text rendering | Improved on V7; V8 reads quoted prompt strings more accurately |
| Access | Web app (midjourney.com) · Discord · alpha.midjourney.com still hosts the rating + experimental surfaces |
| Public API | None generally available · Enterprise API is invitation-gated |
| Commercial rights | Included on all paid tiers · Free users have no commercial rights · $1M+ revenue orgs require Pro or Mega |
| Free tier | None (discontinued late 2024) |
Every data point above was verified against vendor documentation on 2026-05-13. See Sources.
What it actually is
One image-generation product served through two interfaces: a web app (midjourney.com, with alpha.midjourney.com still hosting the rating and experimental surfaces) and a Discord bot. A single subscription covers both.
V7 handles production workloads as the default. V8.1 is now reachable from both Discord and midjourney.com after the April 30 rollout out of alpha-only, with HD temporarily SD-by-default during a server transition. Niji 7 is the sister model for anime and illustrative styles. V8.2 is in training, fed by ongoing community rating data.
Output editing happens through the web editor. Remix, Vary Region, Pan, Zoom Out, Retexture, Paint, Smart Select, and Move/Resize are all first-class tools, not after-the-fact plugins.
The moats: Midjourney’s aesthetic prior (cinematic composition, lighting, color theory) remains the clearest differentiator against every other major generator; the Organize page gives a purpose-built asset library Discord can’t match; and Niji 7 removes the need for a separate anime-specialist tool. The API gap is a real weakness for developers but a deliberate product choice, not a roadmap item.
When to pick Midjourney
- You need the strongest “does this look good” output. Midjourney still tops blind preference tests for cinematic and stylized work. Flux and GPT Image 2 are closer on photoreal than they were a year ago, but not on mood and composition.
- You work on moodboards, concept art, or marketing visuals. The web editor’s Organize page (folders, bulk actions, search across prompts and parameters) is the closest thing to a proper asset library in any image generator.
- You shoot anime or illustrative styles regularly. Niji 7 is bundled in the same subscription. No second tool required.
- You need Stealth mode for client work. Pro and Mega keep generations off the public feed, which matters for agencies and freelance designers under NDA.
- Your shop can live without an API. Solo designers and small teams doing manual work get the most out of Midjourney; the moment workflows need automation, the trade-off shifts.
When to pick something else
- Open weights, self-hosting, or LoRA fine-tuning: Flux (Flux.1 Kontext is the current instruction-editing family from Black Forest Labs, available via API and as open-weight Dev checkpoints).
- Text-in-image reliability (logos, thumbnails, posters): Ideogram still beats V7 here. V8.1 (now broadly accessible) narrows the gap on quoted strings, but Ideogram remains the safer pick for production text work.
- Image gen bundled with a chatbot: GPT Image 2 inside ChatGPT (replaced DALL-E in March 2025). One subscription covers text + image.
- Free tier as a starting point: Leonardo or Stable Diffusion (self-hosted). Midjourney has no free tier.
- Programmatic API for production workflows: Flux, GPT Image 2, or Stable Diffusion. Midjourney’s Enterprise API is invitation-only; unofficial third-party wrappers violate Midjourney’s ToS and risk account bans.
Pricing
Subscription pricing via midjourney.com pricing. All tiers include commercial rights for sole operators; businesses grossing over $1M/year must be on Pro or Mega.
| Plan | Monthly | Annual (effective/mo) | Fast GPU hours | Relax mode | Stealth | Who’s it for |
|---|---|---|---|---|---|---|
| Basic | $10 | $8 (~$96/yr) | 3.3 hrs | No | No | Hobbyists testing the model |
| Standard | $30 | $24 (~$288/yr) | 15 hrs | Unlimited | No | Most working designers land here |
| Pro | $60 | $48 (~$576/yr) | 30 hrs | Unlimited | Yes | Agency or freelance client work |
| Mega | $120 | $96 (~$1,152/yr) | 60 hrs | Unlimited | Yes | Sustained Fast-mode workloads |
Prices verified 2026-05-13 via Midjourney Comparing Plans. Annual billing saves 20% across all tiers. No free trial since late 2024.
Against the alternatives
| Midjourney V7 | Flux.1 Kontext [Pro] | GPT Image 2 | |
|---|---|---|---|
| Photoreal quality | Strong; V8 Alpha stronger still | Strong, more neutral prior | Strong, slightly softer |
| Stylized / cinematic output | Strongest | Mid · tuned for edit-fidelity | Mid |
| Text rendering | Improved on V7; V8 quoted-strings better | Good on instruction edits | Good |
| Commercial rights | All paid tiers (Pro/Mega if org >$1M/yr) | Per-model license (Dev is non-commercial) | Bundled with ChatGPT paid tiers |
| Speed | V7 baseline; V8 Alpha ~4-5× faster | Fast on Pro/Max endpoints | Fast |
| API access | None public | Full public API + open weights on Dev | Via OpenAI API + ChatGPT |
| Best viewed as | Visual-quality specialist | Developer-friendly edit/gen toolkit | Default bundled gen |
Failure modes
- No public API. The official Enterprise API is invitation-only. Third-party wrappers (APIFRAME, ImagineAPI, PiAPI, and others) use browser emulation or bot automation that violates Midjourney’s ToS. Accounts can be banned. If automation is required, pick a different tool.
- Text rendering is “better, not fixed.” V7 improved signage-quality text; V8.1 (now broadly accessible after the April 30 rollout) reads quoted strings more accurately; neither hits Ideogram-grade reliability for logos and dense typography.
- Hand and finger artifacts still surface. V7 reduced the rate materially, but complex poses with multiple hands or interlaced fingers remain a known weakness.
- V8.1 rollout is recent and partly server-throttled. As of April 30, 2026, V8.1 runs on both Discord and midjourney.com, but HD is temporarily SD-by-default during a server transition (use
--hdor settings to opt back in). V8 Alpha-era access on alpha.midjourney.com remains tied to Fast mode for experimental work. - Moderation false-positives. Safety filters reject prompts containing anatomical, violent, or political terms even in obviously benign contexts (medical illustration, historical reference, costume work). Appeals are slow.
- Distinctive “Midjourney look” leaks through. Experienced viewers can identify MJ output by its lighting and composition defaults. Using
--raw, Style References, or Personalization profiles helps; it does not eliminate the signature. - Basic plan images are public by default. Stealth mode (private gallery) requires Pro or Mega. Freelancers on Basic or Standard leak client concepts to the public feed unless they manually manage visibility.
- Revenue threshold for commercial rights. Companies grossing over $1M/year must be on Pro ($60) or Mega ($120). Basic and Standard commercial rights are for sole operators and smaller shops only.
- Copyright on unmodified outputs. Under U.S. law, purely AI-generated images likely have no copyright protection. Midjourney grants usage rights, but you cannot stop others from copying unmodified outputs. Substantive human editing strengthens the claim.
Methodology
This page was produced by the aipedia.wiki editorial pipeline, an automated system that ingests vendor documentation, verifies pricing and model details against primary sources, and generates the editorial analysis you are reading. No individual human wrote this review. Scoring follows the four-dimension rubric at /about/scoring/ (Utility x Value x Moat x Longevity, unweighted average). Last verified 2026-05-13 against Midjourney Version docs, Comparing Plans, Using Images & Videos Commercially, the V8.1 update, the V8 Alpha update, and the Niji V7 announcement.
FAQ
Is Midjourney free? No. The free trial was discontinued in late 2024. The cheapest plan is Basic at $10/month. Free alternatives include Leonardo and self-hosted Stable Diffusion.
What is the current Midjourney version?
V7 is the production default across the web app and Discord as of May 2026. V8.1 graduated out of alpha-only access on April 30, 2026 and is now available on both Discord and midjourney.com, with HD temporarily SD-by-default during a server transition (--hd or settings flips it back). V8.2 is in training, fed by community rating data.
What’s new in V8 and V8.1? V8 Alpha (March 17, 2026) runs roughly 4-5x faster than V7 in Fast mode, reads prompts more accurately, and renders quoted text strings with notably better legibility. V8.1 (April 14 alpha, broadened to Discord and midjourney.com on April 30) sharpens output on SREFs, Moodboards, and HD images and adds native 2K when HD is enabled (V8.1 update, V8 Alpha update).
Does Midjourney have an API? Not a generally-available one. An Enterprise API exists but is invitation-gated via the Midjourney developer application. Unofficial third-party wrappers (APIFRAME, PiAPI, ImagineAPI, and others) operate via browser automation and violate Midjourney’s ToS. They risk account bans and should be treated as unsupported.
Do I get commercial rights to my images? Yes for paid tiers. Basic, Standard, Pro, and Mega all include commercial rights. Free-trial users have no commercial rights. Businesses grossing over $1,000,000/year must be on Pro or Mega to use outputs commercially (Using Images & Videos Commercially).
Web app or Discord, which should I use? Most new users should start on the web app. The editor now carries the heavier toolset (Remix, Vary Region, Pan, Zoom Out, Retexture, Smart Select) and the Organize page handles folders, search, and bulk actions Discord has no equivalent for. Discord still wins on community and live feedback, and the same subscription covers both (Web vs Discord).
Midjourney vs Ideogram for text in images? Ideogram still edges V7 on reliability for dense typography (logos, posters, thumbnail copy). V8.1 (now broadly accessible on Discord and midjourney.com) narrows the gap when prompts enclose target strings in quotation marks. For text-critical production work today, Ideogram remains the safer pick.
Sources
- Midjourney Version docs: current default model, version history
- Comparing Midjourney Plans: pricing and tier features
- V8 Alpha update: V8 Alpha and V8.1 Alpha release notes
- V8.1 update (April 30, 2026): V8.1 graduation out of alpha-only, server-transition HD note, V8.2 training signal
- Niji V7 announcement: anime model release
- Using Images & Videos Commercially: commercial rights and revenue threshold
- Web vs Discord: interface parity and tooling differences
- Midjourney Terms of Service: asset license and content grant
Related
- Category: AI Image Generation
- Comparisons: Midjourney vs Flux · Ideogram vs Midjourney · Adobe Firefly vs Midjourney · Canva AI vs Midjourney
