Last updated: 2026-11-02
What changed
Midjourney is a self-funded AI lab whose latest model, the default since July 2026, turns text prompts into artistic images and short videos. It runs through a web app and Discord, has no free plan or public API, and sells four subscriptions measured in GPU hours rather than per image.
About Midjourney
Midjourney is a generative AI lab founded in 2021 by David Holz that turns text prompts into images and short videos. It has stayed self-funded, and it ran only inside Discord until a web app arrived in 2024; both routes still work today.
The current default model is V8.2, which Midjourney's version notes say became the default on 24 July 2026. It focuses on aesthetics, image quality and Personalization, the feature that learns your taste from images you rate, and it brings an Edit Model that replaces the older Omni Reference, Character Reference and Retexture tools. The release before it, V8.1, made standard jobs four to five times faster than earlier versions and added native 2K HD images without upscaling. Draft mode, style references, moodboards and Remix carry on, and the Niji 7 model, built with Spellbrush, covers anime and illustration styles. On its Alpha site Midjourney is also testing a "Thinking Mode" rerun that it says improves prompt accuracy and typography.
Plans are priced by GPU time rather than image count. Every tier includes fast hours; Standard and above add unlimited Relax generations for images, Pro and Mega extend Relax to SD video, and only Pro and Mega offer Stealth Mode to keep work out of the public gallery. Paid subscribers may use what they make commercially, but a company earning more than $1 million a year must be on Pro or Mega.
Midjourney competes in the AI image generators category with Ideogram for lettering, Adobe Firefly for Creative Cloud users, plus the open-weight FLUX models. General assistants draw images too: ChatGPT does, as does Gemini. For video, Runway is the closer fit. The best AI image generators roundup and the best AI for graphic design guide rank it against the field, and it also appears in Hokai's top AI tools of 2026 list.
Who should look elsewhere: there is no free plan and no official public API, so developers who need to call an image model from code are better served by FLUX or an API-first service. Anyone who needs exact lettering or brand-safe licensing on day one should test Ideogram or Firefly first, and teams that need private work by default pay for the Pro tier just to switch off public visibility.
Pricing
Four plans, all paid: Basic $10/month (3.3 fast GPU hours, or 200 minutes), Standard $30/month (15 hours), Pro $60/month (30 hours) and Mega $120/month (60 hours). Paying a year up front takes 20% off, which works out to $96, $288, $576 and $1,152 a year. Extra fast time costs $4 an hour on every plan.
Standard and above include unlimited Relax image jobs; Pro and Mega also include Relax SD video and Stealth Mode. Companies with more than $1 million in yearly gross revenue must buy Pro or Mega. Source: Midjourney's plan comparison page, read on 2026-11-02.
Before paying, compare the credit-based plans of Ideogram.
| Tier | Monthly price | What it includes |
|---|---|---|
| Basic | $10/mo | 3.3 fast GPU hours a month, SD video, 3 concurrent fast image jobs, General commercial terms |
| Standard | $30/mo | 15 fast GPU hours a month, Unlimited Relax image jobs, SD and HD video, 3 concurrent fast or Relax image jobs |
| Pro | $60/mo | 30 fast GPU hours a month, Unlimited Relax images and SD video, Stealth Mode, 12 concurrent fast image jobs |
| Mega | $120/mo | 60 fast GPU hours a month, Relax Mode covering both stills and SD clips, Private generations kept out of the gallery, Up to 12 fast video jobs at once |
Feature Comparison by Tier
| Feature | Basic | Standard | Pro | Mega |
|---|---|---|---|---|
| Monthly Cost | $10 | $30 | $60 | $120 |
| Fast GPU Hours/Month | 3.3 | 15 | 30 | 60 |
| Relax Mode | — | Unlimited images | Images and SD video | Images and SD video |
| Video resolution | SD | SD and HD | SD and HD | SD and HD |
| Concurrent image jobs | 3 fast | 3 fast or Relax | 12 fast or 3 Relax | 12 fast or 3 Relax |
| Stealth Mode (Private) | — | — | ✓ | ✓ |
| Commercial use | ✓ | ✓ | ✓ incl. companies over $1M | ✓ incl. companies over $1M |
| V8.2 + Edit Model | ✓ | ✓ | ✓ | ✓ |
Key Features
- Text-to-Image Generation: Turns a written prompt into a grid of four images, with aspect ratios up to 14:1 and parameters for chaos, stylize and weirdness.
- V8.2 as the default model: The default since 24 July 2026, V8.2 targets aesthetics and image quality and reads your Personalization profile more closely than V8.1 did.
- Edit Model: One editing model now handles pan, zoom out and region edits, replacing the separate Omni Reference, Character Reference and Retexture tools.
- Native HD images: V8.1 and later render 2048-pixel images without upscaling; an HD job costs 1.3 minutes of GPU time against 0.8 for standard.
- Draft mode and Remix: Draft mode sketches compositions cheaply before a full render, and Remix lets you change the prompt while varying an image you like.
- Image-to-video: Images can be animated into short clips; Relax-mode SD video is unlimited on the Pro and Mega plans.
- Discord & Web Interfaces: Work through midjourney.com or the Discord bot, with a shared gallery, moodboards and style-reference codes on both.
Pros
- Industry-leading artistic quality with exceptional coherence in complex anatomy, lighting, and textures across styles
- Inclusive commercial rights across all paid tiers with no enterprise licensing restrictions for most users
- A large Discord community and public gallery supply prompt ideas, style codes and feedback for new users
- Affordable entry point compared to hiring a human illustrator, with the Standard tier unlocking unlimited Relax Mode generations
- Rapid iteration pace: two new default models shipped in 2026, each tightening prompt following or image quality
Cons
- No free tier since March 2023: a paid subscription is required before you can generate any images
- Getting consistent results still depends on learning parameters and style codes, whether you use the web app or Discord
- Privacy is public by default; keeping work private (Stealth Mode) requires upgrading to the Pro tier
- No official API; third-party Midjourney API integrations exist but carry terms-of-service risk and may violate platform policies
- Ongoing copyright litigation (Disney, Universal, Warner Bros. filed suits in 2025 alleging training data infringement); no SOC2/HIPAA compliance
Product Information
- Cloud
- Yes
- Self-Hosted
- No
- On-Premise
- No
- Languages
- English, Spanish, French, German, Japanese, Chinese, Korean
- Training
- Official Documentation, Video Tutorials, Discord Community, Monthly Webinars, User-Created Guides
Data Handling
- Compliance
- Self-funded (no external investors); no publicly listed SOC 2 · ISO 27001 · HIPAA · or PCI DSS certifications. Data stored on US-based servers. Moderation and content filters in place
Frequently Asked Questions
How much does a Midjourney subscription cost?
Basic is $10 a month, Standard $30, Pro $60 and Mega $120, and paying for a year up front takes 20% off each one. The plans differ mainly in fast GPU hours, from 3.3 to 60 a month, and in whether slow-queue Relax jobs are unlimited. Extra fast time is sold by the hour.
Can you try Midjourney for free?
No. Every way to generate images needs a paid plan, and there is no ongoing free tier or guaranteed trial in 2026. The cheapest route is a single month of Basic, which you can cancel before it renews.
Which image generators compete with Midjourney?
Ideogram is the pick when lettering inside the image has to be right, Adobe Firefly suits teams already on Creative Cloud, and FLUX is the open-weight option for people who want to run or fine-tune a model themselves. ChatGPT and Gemini also make images inside their assistants.
Is Midjourney or Ideogram better for designs with text?
Ideogram has built its reputation on legible typography, so it is the safer first test for posters, logos and social graphics with words in them. Midjourney is the stronger choice for painterly, cinematic or stylised art where mood matters more than exact letters, and its Thinking Mode test targets that typography gap.
What changed in Midjourney in 2026?
An April release brought much faster renders and native 2K images, and a July release became the default with a new Edit Model replacing Omni Reference and Retexture. Niji 7 launched in January for anime styles, and in October Midjourney began testing a Thinking Mode on its Alpha site.
Top Alternatives
- Ideogram: Ideogram is the stronger choice once embedded text has to render correctly inside the image; Midjourney still leads on painterly coherence and anatomy for pure artistic output.
- Adobe Firefly: Teams already living inside Creative Cloud get tighter suite integration from Adobe Firefly; Midjourney rewards faster iteration and draws on a much larger creative community for prompt ideas.
- Runway: Runway is the one to reach for once the deliverable is video rather than a still image; Midjourney stays the deeper tool for image quality and stylistic range on its own.
- FLUX: FLUX suits builders who want an open-weight model they can run and fine-tune themselves; Midjourney keeps the edge on curated artistic output and the tooling built around its own community.
Hokai guides covering Midjourney
- Midjourney vs Nano Banana vs GPT Image: Which Image AI Fits Your Work in 2026?: Midjourney V8.2, Nano Banana 2.1 and GPT Image 2.6 compared on price, blind-vote scores, editing, rights and free access, with sources read on 5 Nov 2026.
- Best AI Tools for Graphic Designers in 2026: Six AI design tools compared by job: Canva, Recraft, Photoroom, Leonardo.Ai, Adobe Firefly and Midjourney, with plans re-read on 8 Nov 2026 so you pick by job.
- Best AI Image Generators in 2026: Pricing, Licensing, and the Real Cost Per Image: Midjourney sells no team plan; Ideogram charges $30 per user and Leonardo $24 a seat. Solo, team and per-image prices for 8 AI image generators, checked.
- What Adobe Firefly Is Actually For, Now It Runs 30 Models It Can't Fully Indemnify: Adobe Firefly ARR is nearing $300 million on freemium growth, and its indemnification promise no longer covers all 30-plus models in its own workspace.
Also mentioned in
- Best AI 3D Model Generators in 2026: What a Credit Really Buys: Six AI 3D generators compared on credits, cost per model and output rights: Tripo, Meshy, Rodin, Hunyuan 3D, Triverse and Spline, priced on 8 November 2026.
- AMD Buys World Labs for $8.2B: What It Means for Marble Users: AMD agreed to buy World Labs for about $8.2B in stock on 28 Sept 2026, and Fei-Fei Li joins AMD. What it means for Marble users before the deal closes.