Want better AI images? I tried Midjourney 7 and it blew me away – here's why

Is it me, or are AI-generated images getting ridiculously good? I swear, every time I scroll, I question whether what I’m seeing is real. It’s not just artists and meme pages using it anymore. It’s everyone.

Don’t lie. You paused today and thought, “Is this even real?” 😩

Well, things just got worse or better… You know what I mean. Midjourney recently announced Midjourney 7, its most advanced model yet, and after seeing what it can do, all I can say is wow.

Also: The dead giveaway that ChatGPT wrote your content – and how to work around it

Today, we are breaking down what makes Midjourney 7 different, how it stacks up against ChatGPT-4o, and the underrated prompts and workflows that can turn it into a secret weapon.

If you are new to my work, my name is Les, but you can call me Mr. Mapp. I kid, I kid; call me Les. I’m a founder with a successful exit and an award-winning performance marketer. These days, I am focused on helping everyday people make sense of all the “AI is changing everything” news.

If you like data-driven marketing insights and want to stay up to date with practical ways to use AI, check out my free newsletter, No Fluff Just Facts. I break down what is actually working in marketing and the AI tools worth paying attention to. 😩

But enough about me. Let’s jump into how Midjourney 7 has me wondering if we live in a simulation.

What Midjourney 7 does differently

If this is your first time hearing about Midjourney, it’s an AI image generation platform that exploded in popularity around 2022. Built inside of Discord, Midjourney made it easy for anyone to create stunning art through simple text prompts, all while building a loyal community. It is often compared to tools like ChatGPT but focuses entirely on visual creativity.

I’ve always found it fascinating that Midjourney used Discord as its primary interface. It was a unique approach that sped up iteration, but what really stood out to me was the community it created. That’s what makes Midjourney special to me. 🫶

Also: Why I just added Gemini 2.5 Pro to the very short list of AI tools I pay for

Fast forward to today, and Midjourney has outdone itself with its latest update, Version 7. When compared to previous models, Version 7 feels like night and day. Getting the output you want is much easier, and some of the problematic prompts that required abstract thinking now work much better.

The image quality is outstanding, and I love how the textures look far more realistic than in previous models.

While this update is great, I still see areas that can be improved. Background faces in crowded scenes still don’t look quite right, and its text generation isn’t on par with ChatGPT (at least not yet).

However, two of the most significant updates in the new model are that personalization is now enabled by default, and a new draft mode has been added, allowing you to generate images in real-time. 🤯

After a quick setup, the system learns your style preferences and tailors outputs accordingly. You can toggle this feature on or off at any time, giving you complete control.

Also: Is that image real or AI? Now Adobe’s got an app for that – here’s how to use it

The updated web interface supports conversational changes and voice input, making it more intuitive. This is helpful for collaborative environments or working on the go.

The updated editor includes a Smart Selection Tool, retexture options, and a new layer system, giving you more precise control over your image adjustments.

At the time of writing, the only way to access the new model is through midjourney.com. Visit the site, click “Create” from the left menu, then click “Settings” and select Version 7 as your default model.

To unlock Version 7, you first need to unlock your personalized style. Unfortunately, using Version 7 without training a new personal style is impossible. Your personalized style from Version 6 is incompatible with Version 7, so you must train a new one.

To do that, click “Unlock Personalization” and quickly rate 200 images. When you return to the creation screen, you will see that Version 7 will be available. 🤓☝️

Want more AI tips? Sign up for Innovation, ZDNET’s weekly newsletter.

How people are using Midjourney 7 (beyond art)

Midjourney 7 powers a range of industries, extending its use beyond traditional digital art.

• Product Mockups: Teams can generate packaging designs, prototypes, and product visuals for pitches or internal reviews, eliminating the need for initial design drafts or photo shoots.

• Pitch Decks & Presentations: Founders can develop compelling visuals for decks, reports, and investor presentations, making their ideas more persuasive.

• Social Media Content: Social media managers can produce branded posts, story templates, and campaign visuals on demand, allowing for quick adaptation to trends or client feedback.

• Education & Training: Educators and trainers can efficiently create visual materials, infographics, and illustrative content for courses and workshops.

• Architecture & Interior Design: The improved spatial understanding and texture rendering make Midjourney 7 useful for visualizing interior layouts, architectural concepts, and environmental designs.

My Midjourney secret hack for getting the best results

When it comes to getting the most out of Midjourney, it comes down to two core principles: 💪

• Knowing what you want
• Being able to communicate what you want

The better you clearly tell the AI what you want, the better your results will be.

Here is my strategy for making the most out of AI image generators:

1. Subject + Setting
2. Elements
3. Scenario + Mood

Here’s what that looks like in practice:

Subject – My muse

I don’t say, “Create an astronaut.” Instead, I say, “Create an astronaut in a cracked golden helmet, worn white space suit covered in dust.”

Set the stage – I always give the subject a world to live in.

Instead of just saying “an astronaut in space,” I provide the context: “an astronaut standing alone on a deserted alien desert at sunset, with broken mountains in the distance and twin suns low in the sky.”

Elements – I stack a few small but important props or features.

Something like “a shattered rover half-buried in the sand, old mission flags flapping weakly, scattered gear around the astronaut’s boots.” These details build the story around the subject.

Scenario + Mood – I ensure some emotion or action is happening.

Maybe it is “reflecting after a failed mission” or “searching for a lost signal.” When you stack these pieces together, the prompt comes out something like this:

“An astronaut in a cracked golden helmet and dusty white space suit, standing alone on an alien desert at sunset, broken mountains and twin suns in the background, a shattered rover half-buried in the sand, old mission flags flapping weakly, the astronaut is searching for a lost signal, somber and reflective mood.”

I’m guiding the AI to what I want rather than leaving the AI to decide. Think “Do it with me, not do it for me.” 

Remember that as the model learns your preferences, it tailors images to you, but the downside is that it can create an echo chamber where everything starts to look the same.

My concerns 

AI has been a major leap for humanity, but much of that progress has come at the expense of creators. They deserve fair compensation, and solutions to rectify this are needed sooner rather than later.

That said, my bigger concern isn’t with Midjourney. It is more about the industry as a whole. ⚠️

Also: The best AI for coding in 2025 (including two new top picks – and what not to use)

It’s getting harder to tell what’s real and what’s AI-generated. That opens up some things we need to consider, like deepfakes, misinformation, and scams becoming harder to spot. News travels faster than ever, and people are slower to double-check what they see. I never thought I would see the day when memes are treated like facts, but here we are.

I am not here to paint a worst-case scenario, but it is clear this tech could cause real problems if it is not handled responsibly. We are past the phase of just talking about it. It is time for smarter safeguards and some actual guidelines to keep things on track.

As AI evolves rapidly, the laws protecting people like you and me must evolve faster.

My two cents

Midjourney 7 is hands-down one of the best image-generation models currently available. The new features, prompting styles, and personalization make it an insanely powerful tool.

I’m all for using this to create better opportunities for ourselves, but if I’m being real, it’s not Skynet that scares me. It’s what people might do with it.

😇 Hope this helps. I’m rooting for you.

P.S. If you want more easy and helpful AI tips and tricks, sign up for my free newsletter, No Fluff Just Facts.

Want more stories about AI? Sign up for Innovation, ZDNET’s weekly newsletter.