Midjourney V6 is out! Is it better than DALL-E 3?
They say that the world is never going to be the same again, people will lose their jobs, advertising will be flipped upside down — Midjourney V6 is here!
Really? Let’s see what the fuss is about…
The All-mighty Midjourney V6 Is Here (ALPHA Version for Now)
Midjourney V6 has emerged as a formidable competitor to DALL-E 3, another prominent player in the text-to-image AI landscape. Both platforms have garnered significant attention for their ability to generate captivating and realistic images from textual descriptions. However, Midjourney V6 stands out with its exceptional capabilities in several key areas.
Midjourney V6 is currently in its Alpha version, which means it is still under development and subject to further improvements. Users may encounter occasional bugs or glitches as the platform continues to evolve.
Midjourney V6 and DALL-E 3, two of the most prominent players in the AI art arena, offer unique strengths and weaknesses. Let’s delve into a comparative analysis to understand their respective advantages:
How to Get Access to Midjourney V6?
With DALL-E 3 accessible in Bing Image Creator for free, or in ChatGPT 4 for 20$ and SDXL which is open source and accessible more and more in graphic design apps like Canva or MockoFun, how can you access Midjourney and how much does it cost?
Midjourney subscription starts at $10/month.
If you already have a Midjourney subscription, you’re probably still on Midjourney V5. To switch to the new improved version, here’s what you need to do:
- Go to the Midjourney Discord server
- Type “/settings” in the chat bar.
- You’ll be prompted with a list of Midjourney versions you can toggle between.
- Your default Midjourney version is likely V5.2, but you can switch to V6 for a limited time.
- Then you can create your prompts as usual just by typing “/imagine” followed by whatever you can dream up.
Developers note that Midjourney V6 requires you to prompt differently than previous versions. V6 is much better at understanding you, so be as explicit in your prompts as you can be. If you want something more realistic, end your prompt with “–style raw” instead of asking with words.
Please note that recently, Midjourney has opened up access to image generation via a website with a web interface, which is much more userfriendly than the interraction with the Discord bot. Currently, access to image generation via the website is only offered to subscribed users that have generated at least 10,000 images.
So, let’s see some examples of images generated with Midjourney V6:
1. Ultra High Quality for Photorealistic Images
Midjourney V6 excels in generating photorealistic images, particularly in the realm of Instagram-like photography. Its images often possess a natural and authentic look, capturing the essence of everyday life with remarkable accuracy.
V5 was already ahead in the area of photorealistic images, and V6 seems to be even better.
2. Text Based Prompts
With the extremely high quality offered by the open-source SDXL and the unrivaled DALL-E 3 when it comes to generation of images with text, Midjourney had to step-up its game and come up with its own version that generates text.
Midjourney V6 outperforms DALL-E 3 in handling text-based prompts. It demonstrates a deeper understanding of text and generates images that faithfully adhere to the user’s instructions, resulting in highly accurate and visually appealing outcomes.
And, I gotta say, seems like V6 has really nailed it:
3. Advertising, Product Mockups & Logo Design
This is a real game changer. With AI getting better and better at producing text, this opens up a huge number of possibilities for making ads, packaging and product mockups and even quick logo design.
Graphic designers — beware!
4. Pop Culture Characters
DALL-E 3 has a clear advantage in generating images of pop culture characters. Its understanding of popular culture references and the ability to depict multiple characters on screen make it the preferred choice for creating images featuring well-known figures.
Here’s how Midjourney V6 handles this prompt:
“Walter White and Jesse Pinkman cooking up blue crystals in a lab selfie taken on iPhone”
Pretty damn good!
5. Copyright Concerns
DALL-E 3 has a stricter policy regarding copyrighted characters, which can limit its use for certain types of images. Midjourney V6, on the other hand, offers more freedom in this regard, allowing users to generate images of copyrighted characters without the risk of copyright infringement.
I’ve bumped into the safety and/or copyright restrictions of DALL-E 3 a lot, sometimes leaving me scratch my head as to what I’ve done wrong. These restrictions are a good thing, especially with the growing concerns over photojournalism and ethics regarding fake news.
Not with Midjourney V6 though — haven’t run into any copyright restrictions with characters or brands.
Midjourney V6: Limitations and Future Potential
While Midjourney V6 represents a significant leap forward in AI art, it still has some limitations that hinder its full potential:
Multiple characters: When working with the new Midjourney V6, we noticed it still struggles with generating images featuring multiple characters on screen. The interactions between characters can sometimes appear unnatural or disjointed, requiring further refinement.
Upscaling: Midjourney V6’s upscaling capabilities, while impressive, can sometimes introduce artifacts or distortions into the generated images. Further improvements in upscaling algorithms are needed to ensure consistent high-quality results.
Zoom/outpainting: There is no ‘Zoom Out’/outpainting option at the moment. No custom Zoom either.
Conclusion
Midjourney V6 has emerged as a formidable player in the AI art landscape, challenging the dominance of DALL-E 3 with its exceptional image quality, enhanced text accuracy, and community-driven development approach. While it still has some limitations, Midjourney V6’s potential is immense, and it is poised to revolutionize the way we create and interact with AI-generated art. As the platform continues to evolve and improve, we can expect even more groundbreaking developments in the realm of AI art.
