| Plan | Monthly | Best for |
|---|---|---|
| Basic | $10 | Casual use, occasional generation |
| Standard | $30 | Regular creator work, most subscribers |
| Pro | $60 | Heavy creator use, agencies, freelancers |
| Mega | $120 | Studios, high-volume production |
Standard v8 generations are included in your subscription credits. Two modifiers burn through credits faster:
- The
--hdflag uses 4x GPU time per generation. - The
--q 4flag (max quality) uses 4x GPU time. - Combined (–hd –q 4) the two stack to 16x GPU time.
For a Mega user pulling HD heavy outputs on every generation, this means burning the monthly allowance roughly 16 times faster than standard generation. For most creators, leaving v8 on default settings is the right balance.
How do you use Midjourney v8?
Three entry points, in order of how the community uses the tool:
Discord. Sign up at midjourney.com, link a Discord account, join the Midjourney server. Use the /imagine command followed by your prompt. The output appears in the channel, visible to other users in that channel (a feature, not a bug — the Discord experience is collaborative on purpose). For private generation, use Midjourney’s bot in your own server or DM the bot directly. This is where the culture is, and where most heavy users still spend their time.
The web app. Available at midjourney.com/app for paid subscribers. Cleaner interface, no Discord required. The web app is becoming the default entry point for new users in 2026, especially those who do not already use Discord. Functionality is at parity with Discord for image generation; video features rolled out to the web app first.
The API. Midjourney’s official API is still in limited rollout as of May 2026. For now, most programmatic Midjourney workflows route through third-party API wrappers (which the company tolerates but does not endorse) or through the web app’s scripting capabilities.
Three prompting habits worth picking up:
- Use quote marks for text. “A vintage poster of a city skyline with the text “NEW YORK” in bold serif” produces readable text in v8 where earlier versions would have garbled it.
- Reference real artists and aesthetics. Midjourney’s training emphasizes art history and design references more than most competitors. Specifying “in the style of Edward Hopper” or “Wes Anderson color palette” lands more reliably here than in other models.
- Use the parameter flags sparingly.
--ar 16:9for aspect ratio is essential.--hdand--q 4are worth saving for shots you actually want to push to maximum quality.
How does Midjourney v8 compare to other image models?
| Model | Strength | Best for |
|---|---|---|
| Midjourney v8 | Stylistic range, aesthetic polish | Single hero images, mood boards, art-direction-heavy work |
| Nano Banana 2 | Identity preservation, web grounding, text rendering | Multi-panel sequences, storyboards, brand consistency |
| DALL-E 3 | ChatGPT integration | Casual generation inside ChatGPT workflows |
| Stable Diffusion (open-weights) | Local control, custom fine-tuning | On-prem workflows, custom models |
| Flux (Black Forest Labs) | Photoreal output, portrait quality | Realistic portraits, product photography |
There is no head-to-head winner overall. The honest framing: Midjourney still has the most distinct aesthetic of any major image model — the “Midjourney look” is recognizable. For art-directed work where style is the unlock, it is hard to beat. For consistency across a sequence, Nano Banana 2 is stronger. For local control, open-weights Stable Diffusion is the choice.
What are Midjourney v8’s limitations?
Four practical limitations.
No identity preservation across images. Generate the same character twice and you get two different people. Midjourney has the cref (character reference) and sref (style reference) features which help, but they do not match Nano Banana 2’s identity-locking capability for sequential work. If you need the same protagonist across a 12-panel comic, Midjourney is the wrong tool today.
Discord-first culture is divisive. Half the audience loves it (the community, the public gallery, the collaborative pace). The other half finds the Discord interface confusing, the public visibility uncomfortable, or the bot-command syntax fiddly. The web app helps, but the cultural center of gravity is still Discord.
Limited API maturity. If you are building Midjourney into a product, the official API is still in limited rollout. Third-party wrappers exist but are not endorsed and can break when Midjourney updates the underlying interface.
Video is early. V8 video at 10 seconds and 60fps is impressive, but it is one generation behind the dedicated video labs. For serious video work, Veo 3.1, Runway Gen-4.5, Kling 3.0, and Seedance 2.0 are all stronger choices. Midjourney’s strength remains stills.
Get Smarter About AI Every Morning
Free daily newsletter — one story, one tool, one tip. Plain English, no jargon.
Free forever. Unsubscribe anytime.
Frequently Asked Questions
Do I have to use Discord to use Midjourney?
No. The web app at midjourney.com/app gives you full access to v8 image generation. Most new users in 2026 start with the web app. Discord remains the cultural center, but it is no longer required.
Is there a free trial?
Not currently. Midjourney removed its free-trial program in 2023 due to abuse and has not reinstated it. The cheapest entry point is the Basic plan at $10/month.
Can I use Midjourney images commercially?
Yes, on paid plans. Midjourney’s terms permit commercial use of generated content on the Basic, Standard, Pro, and Mega tiers. Free-trial output (when it existed) and any output flagged for moderation issues are excluded.
What about Midjourney v9?
No public roadmap as of May 2026. Midjourney’s historical cadence has been roughly one major version per year, so v9 is plausible in late 2026 or early 2027. The company does not pre-announce releases.
How does Midjourney’s video compare to Veo or Runway?
Midjourney v8 video is good for stylized 10-second clips at 60fps. Veo 3.1 and Runway Gen-4.5 are stronger across most production-grade video use cases, with native audio, longer clip lengths, and deeper editing tools. If you are doing serious video work, the dedicated video tools win. Midjourney video is a nice addition for creators already inside the Midjourney ecosystem.
What is the “Midjourney look”?
The aesthetic Midjourney’s training and rendering tends to produce by default — warm tones, cinematic lighting, painterly textures, slightly stylized realism. Critics call it formulaic; fans call it polished. V8 expanded the stylistic range, but the default look is still recognizable enough that experienced viewers can often guess which model generated an image.
Sources
- Midjourney Docs — Version reference (including v8 / v8.1)
- Midjourney — Updates and release notes
- WaveSpeed — Midjourney v8 features, pricing, and how to use
You might also like
- Nano Banana 2 Explained — the identity-preserving image model for multi-panel work.
- Google Veo 3.1 Explained — the leading AI video model.
- Runway Gen-4.5 Explained — the #1-ranked video model.
- Kling 3.0 Explained — strongest character continuity in video.
- Seedance 2.0 Explained — cheapest per-second video pricing.
- The Real State of AI Video in 2026 — our Special Report Vol. 2.
- Every AI Model Worth Knowing in 2026.
- AI Glossary — every term in plain English.
AI Summary
The headline: Midjourney v8.1 became the default model on April 30, 2026. Native 2K HD output, 4-5x faster than v7, much better text rendering, and a new 10-second video mode at 60 frames per second.
Who built it: Midjourney — the bootstrapped image-AI company founded by David Holz. Roughly $500M ARR with no venture capital and no marketing budget. 21 million creators in the Discord community.
Best if: You want the best stylistic image output and you do not mind a Discord-first interface (a web app exists too, but Discord is where the culture lives).
Skip if: You need identity preservation across multiple images (use Nano Banana 2), heavy text-on-image work in production (Nano Banana 2 is better), or you cannot tolerate a Discord workflow.
What is Midjourney v8?
Midjourney v8 is the eighth major version of Midjourney’s image-generation model. V8 Alpha launched on March 17, 2026. V8.1 became the default model on April 30, 2026 — the version most users are on as of May.
The headline shift in v8 is native 2K HD output by default. Previous Midjourney versions generated at lower resolution and asked you to upscale separately. V8 renders directly at 2K and is roughly four to five times faster than v7 in the process. For users moving from v7 to v8, the practical experience is markedly snappier prompts and ready-to-print output without a separate upscale step.
Beyond images, v8 introduced native text-to-video and image-to-video at up to 10 seconds at 60 frames per second. This is separate from Midjourney’s older V1 video model, which extends 5-second clips up to 21 seconds — v8 video is a different pipeline focused on shorter, higher-frame-rate output.
Midjourney is a closed-weights foundation model family. It runs on Midjourney’s own infrastructure, accessed through Discord, the web app, or the API. There is no downloadable version.
Who is Midjourney, and how did they get here?
Midjourney was founded by David Holz — previously a co-founder of Leap Motion — and incorporated as an independent research lab in 2021. The company shipped its first public version on Discord in July 2022, ahead of Stable Diffusion’s open release in August 2022 and DALL-E 2’s general availability in September of the same year.
What sets Midjourney apart from Runway, OpenAI, and most other AI image companies is the financial story:
- No venture capital. Midjourney has never raised an institutional funding round.
- No marketing. No paid ads, no growth team, no celebrity partnerships. Growth has been entirely organic through the Discord community.
- ~$500M ARR. Per the most recent Product Growth analysis, Midjourney’s subscription revenue puts it among the highest-revenue private AI companies in the world.
- ~21 million users. Across the Discord community plus the web app.
Holz’s framing for the company has stayed consistent across interviews:
We’re trying to expand the imaginative powers of the human species. The goal is to make humans more imaginative, not make imaginative machines.
David Holz, founder, Midjourney
On the Discord-first interface choice (which has confused outside analysts for years), Holz famously responded: “People like it, and nothing else matters.” The community is the product as much as the model is.
What makes v8 different from v7?
Four upgrades stand out:
Native 2K HD output. No upscale step. The model generates directly at 2K resolution. For poster design, print mockups, and any workflow where the final asset needs to be a high-resolution file, this is a meaningful workflow improvement.
4-5x faster generation. Same prompt, same quality target, far less waiting. Combined with the resolution improvement, this means v8 produces print-ready output in the time v7 produced screen-resolution drafts.
Readable text inside images. Image AI has been historically bad at text — deformed letters, misspellings, font drift. V8 adds a quote-mark syntax that produces readable typography for posters, labels, and signage. Wrap the text you want in double quotes within your prompt and the model treats it as text-to-render rather than text-to-interpret.
Improved spatial composition and material fidelity. Lighting, fabric, water, metal, glass — the things image models historically struggled to render coherently — are more accurate in v8. The model also handles prompt fidelity better, meaning it follows your prompt more literally instead of taking creative liberties.
How much does Midjourney v8 cost?
Midjourney’s subscription tiers have stayed unchanged through the v8 launch:
| Plan | Monthly | Best for |
|---|---|---|
| Basic | $10 | Casual use, occasional generation |
| Standard | $30 | Regular creator work, most subscribers |
| Pro | $60 | Heavy creator use, agencies, freelancers |
| Mega | $120 | Studios, high-volume production |
Standard v8 generations are included in your subscription credits. Two modifiers burn through credits faster:
- The
--hdflag uses 4x GPU time per generation. - The
--q 4flag (max quality) uses 4x GPU time. - Combined (–hd –q 4) the two stack to 16x GPU time.
For a Mega user pulling HD heavy outputs on every generation, this means burning the monthly allowance roughly 16 times faster than standard generation. For most creators, leaving v8 on default settings is the right balance.
How do you use Midjourney v8?
Three entry points, in order of how the community uses the tool:
Discord. Sign up at midjourney.com, link a Discord account, join the Midjourney server. Use the /imagine command followed by your prompt. The output appears in the channel, visible to other users in that channel (a feature, not a bug — the Discord experience is collaborative on purpose). For private generation, use Midjourney’s bot in your own server or DM the bot directly. This is where the culture is, and where most heavy users still spend their time.
The web app. Available at midjourney.com/app for paid subscribers. Cleaner interface, no Discord required. The web app is becoming the default entry point for new users in 2026, especially those who do not already use Discord. Functionality is at parity with Discord for image generation; video features rolled out to the web app first.
The API. Midjourney’s official API is still in limited rollout as of May 2026. For now, most programmatic Midjourney workflows route through third-party API wrappers (which the company tolerates but does not endorse) or through the web app’s scripting capabilities.
Three prompting habits worth picking up:
- Use quote marks for text. “A vintage poster of a city skyline with the text “NEW YORK” in bold serif” produces readable text in v8 where earlier versions would have garbled it.
- Reference real artists and aesthetics. Midjourney’s training emphasizes art history and design references more than most competitors. Specifying “in the style of Edward Hopper” or “Wes Anderson color palette” lands more reliably here than in other models.
- Use the parameter flags sparingly.
--ar 16:9for aspect ratio is essential.--hdand--q 4are worth saving for shots you actually want to push to maximum quality.
How does Midjourney v8 compare to other image models?
| Model | Strength | Best for |
|---|---|---|
| Midjourney v8 | Stylistic range, aesthetic polish | Single hero images, mood boards, art-direction-heavy work |
| Nano Banana 2 | Identity preservation, web grounding, text rendering | Multi-panel sequences, storyboards, brand consistency |
| DALL-E 3 | ChatGPT integration | Casual generation inside ChatGPT workflows |
| Stable Diffusion (open-weights) | Local control, custom fine-tuning | On-prem workflows, custom models |
| Flux (Black Forest Labs) | Photoreal output, portrait quality | Realistic portraits, product photography |
There is no head-to-head winner overall. The honest framing: Midjourney still has the most distinct aesthetic of any major image model — the “Midjourney look” is recognizable. For art-directed work where style is the unlock, it is hard to beat. For consistency across a sequence, Nano Banana 2 is stronger. For local control, open-weights Stable Diffusion is the choice.
What are Midjourney v8’s limitations?
Four practical limitations.
No identity preservation across images. Generate the same character twice and you get two different people. Midjourney has the cref (character reference) and sref (style reference) features which help, but they do not match Nano Banana 2’s identity-locking capability for sequential work. If you need the same protagonist across a 12-panel comic, Midjourney is the wrong tool today.
Discord-first culture is divisive. Half the audience loves it (the community, the public gallery, the collaborative pace). The other half finds the Discord interface confusing, the public visibility uncomfortable, or the bot-command syntax fiddly. The web app helps, but the cultural center of gravity is still Discord.
Limited API maturity. If you are building Midjourney into a product, the official API is still in limited rollout. Third-party wrappers exist but are not endorsed and can break when Midjourney updates the underlying interface.
Video is early. V8 video at 10 seconds and 60fps is impressive, but it is one generation behind the dedicated video labs. For serious video work, Veo 3.1, Runway Gen-4.5, Kling 3.0, and Seedance 2.0 are all stronger choices. Midjourney’s strength remains stills.
Get Smarter About AI Every Morning
Free daily newsletter — one story, one tool, one tip. Plain English, no jargon.
Free forever. Unsubscribe anytime.
Frequently Asked Questions
Do I have to use Discord to use Midjourney?
No. The web app at midjourney.com/app gives you full access to v8 image generation. Most new users in 2026 start with the web app. Discord remains the cultural center, but it is no longer required.
Is there a free trial?
Not currently. Midjourney removed its free-trial program in 2023 due to abuse and has not reinstated it. The cheapest entry point is the Basic plan at $10/month.
Can I use Midjourney images commercially?
Yes, on paid plans. Midjourney’s terms permit commercial use of generated content on the Basic, Standard, Pro, and Mega tiers. Free-trial output (when it existed) and any output flagged for moderation issues are excluded.
What about Midjourney v9?
No public roadmap as of May 2026. Midjourney’s historical cadence has been roughly one major version per year, so v9 is plausible in late 2026 or early 2027. The company does not pre-announce releases.
How does Midjourney’s video compare to Veo or Runway?
Midjourney v8 video is good for stylized 10-second clips at 60fps. Veo 3.1 and Runway Gen-4.5 are stronger across most production-grade video use cases, with native audio, longer clip lengths, and deeper editing tools. If you are doing serious video work, the dedicated video tools win. Midjourney video is a nice addition for creators already inside the Midjourney ecosystem.
What is the “Midjourney look”?
The aesthetic Midjourney’s training and rendering tends to produce by default — warm tones, cinematic lighting, painterly textures, slightly stylized realism. Critics call it formulaic; fans call it polished. V8 expanded the stylistic range, but the default look is still recognizable enough that experienced viewers can often guess which model generated an image.
Sources
- Midjourney Docs — Version reference (including v8 / v8.1)
- Midjourney — Updates and release notes
- WaveSpeed — Midjourney v8 features, pricing, and how to use
You might also like
- Nano Banana 2 Explained — the identity-preserving image model for multi-panel work.
- Google Veo 3.1 Explained — the leading AI video model.
- Runway Gen-4.5 Explained — the #1-ranked video model.
- Kling 3.0 Explained — strongest character continuity in video.
- Seedance 2.0 Explained — cheapest per-second video pricing.
- The Real State of AI Video in 2026 — our Special Report Vol. 2.
- Every AI Model Worth Knowing in 2026.
- AI Glossary — every term in plain English.