Midjourney is still the benchmark for artistic image quality, and V8.2 has been the default model since July 24, 2026 — a rewritten codebase producing native 2K output around five times faster than V7. The pricing detail that trips up new subscribers: you're buying fast GPU time, not a number of images. Basic at $10 gives roughly 3.3 hours, about 200 generations, with no Relax mode when it runs out. Standard at $30 adds unlimited slow generation, which is why most regular users end up there.
Reviewed by EdGrows·Updated Aug 18, 2026Affiliate link — we may earn a commission
This analysis is based on documentation, public user reports, and vendor materials — not yet on our own hands-on testing. How we rate
This is the thing almost every new subscriber gets wrong, and it explains most of the complaints about Midjourney pricing.
The plans don't include a number of images. They include fast GPU hours. Basic gives roughly 3.3 hours, which comes out around 200 generations depending on what you ask for — a dense, complex prompt eats more of the clock than a simple one.
When those hours run out on Basic, you stop. There's no Relax mode at that tier, so the options are waiting for the next billing cycle or buying extra hours at about $4 each.
That single fact is why so many people describe Basic as a trial rather than a plan, and why the real recommendation for anyone generating regularly starts at Standard.
| Plan | Monthly | Annual | Fast GPU | The thing that matters |
|---|---|---|---|---|
| Basic | $10 | $8 | ~3.3 hrs (~200 images) | No Relax mode — you stop when it runs out |
| Standard | $30 | $24 | ~15 hrs | Unlimited Relax — you keep going, slowly |
| Pro | $60 | $48 | ~30 hrs | Stealth Mode — your images stay private |
| Mega | $120 | $96 | ~60 hrs | Everything, more of it |
Two upgrades carry real weight and the rest is volume.
Basic to Standard buys Relax mode. Unlimited slow generation with queue waits of roughly zero to thirty minutes. For anyone who generates most days, this converts "I've run out" into "I'll wait a bit," which is a different product.
Standard to Pro buys privacy. Stealth Mode keeps your generations off the public Explore gallery, and it's exclusive to $60 and above. Everything made on Basic or Standard is publicly visible by default.
That second one deserves emphasis because it's rarely in comparison tables: if you do client work, unreleased brand concepts, or anything under NDA, your entry price is $60, not $10. No amount of tier-shopping changes that.
One more rule worth knowing: companies with revenue over a million dollars a year are required to be on Pro or Mega regardless of usage.
The V8 series arrived as Alpha on March 17, 2026, built on a completely rewritten codebase — the most significant architectural change since V5.
What it delivered: roughly five times faster generation, native 2K HD output, substantially better text rendering, and noticeably stronger handling of prompts with multiple elements competing for attention.
V8.1 became the default on June 11, and V8.2 took over on July 24, 2026, refining aesthetics and personalisation further. V7 stays selectable while a few features finish being ported across.
The practical upshot: complaints about Midjourney's weak in-image text are now half-true rather than fully true. It improved a lot in V8 and still trails Ideogram, which is purpose-built for that job. If your work involves posters, packaging or anything with words in it, that gap still decides the tool.
Aesthetic coherence. It's the least measurable thing in this category and the reason Midjourney keeps its position despite being subscription-only with no free tier.
Ask a dozen generators for the same moody cinematic portrait and Midjourney's version will most often be the one that looks like someone made a decision — about lighting, palette, composition — rather than averaging a training set. For concept art, mood boards, editorial illustration and brand visuals, that's the whole job.
The trade is literalism. Midjourney interprets rather than transcribes. Give it a precise spec with exact placement and it will produce something more beautiful than what you asked for, which is delightful in ideation and infuriating in production.
Text-heavy design. Better in V8, still not the specialist. Ideogram renders legible type at accuracy Midjourney doesn't approach.
Free experimentation. There is no free tier and no trial. Leonardo AI and Krea both let you test without a card.
Exact prompt compliance. When a client needs the logo bottom-left at a specified size, a generator that interprets is the wrong tool. GPT Image 2 in ChatGPT is more obedient.
Custom model training. No LoRAs, no fine-tunes, no ControlNet. That's Stable Diffusion and FLUX territory, and it's a hard limit rather than a gap.
Against Ideogram: Ideogram exists because generators couldn't spell, and independent testing still puts its in-image text accuracy around 90% against Midjourney's roughly 30% on short phrases. Midjourney wins comprehensively on photorealism and style. If words appear in the image, Ideogram; if they don't, Midjourney.
Against FLUX: FLUX from Black Forest Labs — founded by the people who built the original Stable Diffusion — leads the open-weight field and can be self-hosted or accessed cheaply by API. It gives you control and no subscription; Midjourney gives you a house aesthetic and no setup. Developers building pipelines take FLUX; art directors take Midjourney.
Against Leonardo AI: Leonardo offers a usable free tier, custom model training and a more conventional interface. It's the pragmatic pick for game asset work and anyone who needs consistency across a set. Midjourney's individual images generally look better; Leonardo's workflow is more controllable.
Against Krea: Krea aggregates dozens of models including FLUX and Ideogram behind one subscription, plus real-time canvas generation. It's the right answer for exploration across many models; Midjourney is the right answer when you already know the look you want.
Against Stable Diffusion: Stable Diffusion is free, self-hostable, and infinitely modifiable through LoRAs and ControlNet. It also demands a GPU and a weekend. Midjourney is a product; SD is a toolkit.
| Plan | Monthly | Annual (20% off) | Fast GPU hours | Relax | Stealth |
|---|---|---|---|---|---|
| Basic | $10 | $8 | ~3.3 | No | No |
| Standard | $30 | $24 | ~15 | Unlimited | No |
| Pro | $60 | $48 | ~30 | Unlimited | Yes |
| Mega | $120 | $96 | ~60 | Unlimited | Yes |
Checked August 2026 against Midjourney's published plans and current third-party verification. Pricing has been stable through 2026 while the model advanced from V7 to V8.2 (default since July 24, 2026). Billing is by fast GPU time, not image count; extra fast hours are around $4 each. No free plan or trial exists. Businesses over $1M annual revenue must be on Pro or Mega. Confirm on midjourney.com before subscribing.
Start at Standard if you generate more than occasionally. The $20 difference buys unlimited Relax generation, which changes Midjourney from a metered resource into something you can use freely. Most people who start on Basic upgrade within two months.
Budget Pro if privacy matters at all. Client work, unreleased products, anything under NDA — Stealth Mode is $60 and there's no way around it.
Avoid Turbo unless you need speed specifically. It runs up to 4x faster and consumes 2x the GPU time, which is a bad trade for routine work.
Skip video unless you're on Pro or Mega. Image-to-video draws from the same GPU allowance with no separate pricing, so it drains a Basic or Standard budget quickly.
Midjourney has held the aesthetic crown through three years of aggressive competition, and V8 explains why it's still holding it. A rewritten codebase delivering native 2K, roughly five times the speed of V7, and text rendering that finally isn't embarrassing — plus commercial rights on every paid tier that survive cancellation, which is a genuinely generous term in this market. For concept art, editorial illustration and brand visuals, nothing else consistently produces images that look this deliberate.
The pricing is where care is needed. You buy GPU hours rather than images, and Basic's roughly 3.3 hours run out around 200 generations with no Relax mode to fall back on. Privacy costs $60 — Stealth Mode is Pro and above, so anything on Basic or Standard is publicly visible on the Explore gallery. There's no free tier and no trial, which makes evaluation a $10 commitment. And Midjourney interprets rather than obeys, which is exactly what you want in ideation and exactly what you don't want when a spec has to be met.
The honest counterpoint to the price complaints: at $30 with unlimited Relax generation, Midjourney is cheaper per usable image than most metered competitors once you account for how many generations you'd otherwise discard.
For anyone whose work is judged on how images look, recommend at Standard — or Pro if the work is confidential. For text-heavy design or precise briefs, use something else and come back for the hero shot.
Note: Midjourney does not currently have an affiliate program with AIVario. We earn no commission, and this rating carries no commercial incentive.
Best for: Concept artists and illustrators, brand and editorial visuals, mood boards and ideation, anyone prioritising aesthetic quality over literal prompt compliance Not ideal for: Text-heavy design work, exact-spec client briefs, free experimentation, custom model training and fine-tuning Bottom line: Still the best-looking output in the category, sold by GPU hour rather than by image — start at Standard, and budget Pro if your work can't be public.
Four plans: Basic at $10 a month, Standard at $30, Pro at $60 and Mega at $120, each 20% cheaper billed annually — $8, $24, $48 and $96 respectively. There is no free plan and no free trial, so the cheapest entry point is $10. Companies earning over a million dollars a year are required to be on Pro or Mega. Annual billing charges the full year upfront.
They're what you're actually buying. Midjourney bills fast GPU time rather than a quota of images, so a complex generation costs more of your allowance than a simple one. Basic includes roughly 3.3 fast hours, which works out to around 200 images. Turbo mode runs up to four times faster but consumes double the GPU time, and once fast hours are gone on Basic you either wait for the next cycle or buy extra hours at about $4 each.
Unlimited slow generation, available on Standard, Pro and Mega — not Basic. Images go into a queue with waits typically running anywhere from nothing to about 30 minutes. This is the single biggest reason to choose Standard at $30 over Basic at $10: once your fast hours run out on Standard you keep generating, just slower, while Basic simply stops. For most regular users that difference is worth the extra $20.
V8.2, default since July 24, 2026. The V8 series launched as Alpha on March 17, 2026 on a completely rewritten codebase, bringing roughly five times faster generation, native 2K HD output, much better text rendering and stronger handling of complex multi-element prompts. V8.1 became default on June 11 before V8.2 replaced it. V7 remains selectable for the handful of features still being ported across.
Only on Pro and Mega. Stealth Mode, which keeps generations off Midjourney's public Explore gallery, is exclusive to the $60 and $120 tiers. Anything generated on Basic or Standard appears publicly by default. For client work under NDA, brand concepts before launch or anything commercially sensitive, that puts your real entry price at $60 rather than $10 — a detail worth knowing before you subscribe.
Better looking, worse at following instructions literally. Midjourney has stronger aesthetic coherence, lighting and style consistency, which is why it dominates concept art and brand visuals. GPT Image 2 in ChatGPT — DALL-E 3 was retired in May 2026 — handles precise prompt interpretation and text inside images better, and comes bundled with a subscription you may already have. Many people use both for different jobs.
Image-to-video only, not text-to-video. A still can be animated into a five-second clip and extended up to 21 seconds, on all plans. The cost consideration is that video draws from the same fast GPU allocation as images rather than having separate pricing, so it burns your monthly budget considerably faster. In practice it's only comfortable on Pro or Mega where the GPU allowance is large enough to absorb it.
Yes, on every paid tier, and they persist after the subscription ends. That's a genuinely favourable term compared with tools that gate commercial usage behind higher plans or revoke rights on cancellation. The caveat is the privacy question above rather than the licensing one: on Basic and Standard your commercially usable images are also publicly visible.
The leading AI voice platform — text-to-speech, voice cloning, and conversational AI in 70+ languages.
The design suite everyone already uses, now bundled with a free professional toolset — and a credit system worth understanding first.
Category-defining design tool with unusually smart seat pricing — provided you assign the right seat type to each person.