Midjourney V7 is Here! Not Just an Image Quality Upgrade, This Time AI Drawing Wants to Read Your Mind

The giant of the AI drawing world, Midjourney, has finally launched the V7 Alpha version! This time, it’s not just about pursuing more beautiful images and smoother logic, but also adding the super cool “Draft Mode” and a default-enabled “Personalization” feature. Come and see the highlights of V7 and how it compares to other AIs!

Hey, fellow AI drawing enthusiasts, have you felt a stir in the Midjourney community lately? That’s right, Midjourney, which always brings visual surprises, finally officially launched the Alpha test version of the V7 image model on April 3rd after much anticipation! As soon as this news came out, it was like dropping a stone into a calm lake, causing quite a stir.

By the way, why did V7 suddenly appear?

In fact, the day before the official debut of V7 (April 2nd), the Midjourney team announced on the official channel that they would end the months-long “RELAX-ATHON” (that is, the benefit period of unlimited use of Relax mode) in order to adjust the server to welcome V7. As soon as this news came out, the community exploded, and everyone was guessing whether the new version was really coming. Sure enough, the next day, the official confirmed the news of the V7 Alpha test and invited everyone to try it out.

Counting the days, it’s been almost a year since Midjourney’s last major version update. Since David Holtz founded Midjourney in 2022, it has made a name for itself in the field of AI drawing with its unique artistic style and high-quality image output. It is said that the development of V7 took several months, adopted a “completely different architecture”, and integrated new data sets and natural language processing technology. The goal is very clear: to make a great leap forward in image quality and understanding of user prompts!

So what are the eye-catching new things in V7?

According to official statements and feedback from netizens who have already played with it, V7 has brought a lot of good things this time:

“Draft Mode”: An accelerator for brainstorming?

This time, V7 has added a function called “Draft Mode”. The team describes it as “probably the best way to iterate on ideas”. Sounds a bit mysterious? Simply put, this mode has a 10 times faster generation speed, but the cost is reduced by 50%! It even supports voice input, allowing your inspiration to be transformed into visual drafts in real time like a “flowing dream”. Imagine that the ideas in your head are almost synchronously turned into pictures, isn’t that cool? For designers or artists who need to quickly brainstorm and try different styles, this function is simply a godsend.

Midjourney Draft Mode video

A great leap in image quality and coherence, are the fingers counted correctly this time?

Having seen some of the V7 generated images shared by netizens, the details are really amazing. The new model has made significant progress in handling complex elements such as human bodies, hands, and object coherence. It is no longer the embarrassing situation of “extra fingers” or “twisted limbs” as before (although more tests may be needed to verify this). At the same time, the image texture is richer and the resolution is higher. It seems that Midjourney has put a lot of effort into “drawing realistically” this time.

Personalization is enabled by default: V7 understands your “beauty” better

This is very interesting! In previous versions, the personalization function might need to be manually enabled or adjusted. But V7 has the personalization function enabled by default. You need to rate about 200 images first to create your personal visual preference profile. After that, the model will adjust the generation results according to your preferences. This makes V7 the first model of Midjourney to have personalization enabled by default, which also means that the images you generate will have more of your own flavor. Unique, sounds very attractive!

Future update plans: more goodies to come

Midjourney has also made a big promise, announcing that in the next 60 days, more new features will be released one after another, including a brand new character and object reference system. These updates will make V7’s capabilities more complete and better meet the needs of professional creation. It seems that V7 is just the beginning, and its subsequent development is worth looking forward to.

Want to try it out? How to play with V7?

Are you tempted? If you want to try the power of V7, you need to complete the “personalization settings” mentioned earlier (rate 200 images). After completion, you can enable the V7 model in Midjourney’s website or Discord chatbot. On the website, there is usually a drop-down menu for version selection, and you can switch quickly with a click.

But don’t worry, the official also reminded that some standard functions of V7, such as Upscaling and Re-texturing, are not yet fully online. It is expected to be launched gradually in the next two months. Therefore, the current Alpha version can be regarded as a trial version for experiencing core functions and new styles.

Not just drawing, Midjourney’s ambition?

The launch of Midjourney V7 not only consolidates its position in the field of AI image generation, but also sets a new benchmark for the entire industry. The market competition is so fierce now, with OpenAI’s DALL-E, Google’s Imagen, and the open source Stable Diffusion, etc. Midjourney seems to have chosen a path that focuses more on image quality, user control, and professional applications, trying to differentiate itself.

What’s more interesting is that Midjourney also announced the formation of a new hardware team and is training video and 3D object generation models! This reveals that they are not just satisfied with being a “drawing tool”, but want to show their skills in the broader field of multimedia creation. This is a big game of chess!

What does this mean for us creators? More powerful tools, more possibilities! From quickly generating conceptual prototypes to generating commercial-grade high-quality images, Midjourney V7 is promoting the deep integration of AI and human imagination.

V7 is amazing, but… how does it compare to GPT-4o?

After talking so much about the good things about V7, some people may ask: “So how does it compare to the recently popular GPT-4o image generation?”

This is a good question. First of all, Midjourney requires a paid subscription. And OpenAI’s GPT-4o, although it has the function of generating images, and is currently available for free in ChatGPT (but with a terrible rate limit, about three times a day, and the generation speed is slow), the strengths of the two are not the same.

According to some preliminary comparisons and user feedback:

• Style and aesthetics: Midjourney V7 is generally considered to be the strongest in shaping specific artistic styles and overall aesthetics. If you are pursuing that unique, exquisite, and artistic picture, V7 may be the first choice.
• Accuracy and comprehension: GPT-4o may perform better in understanding natural language instructions and generating content that is closer to the text description. If what you need is “to draw exactly what I say”, GPT-4o may be more competent.
• Image modification: Google’s Gemini may have an advantage in image modification.
• Price: V7 is paid, GPT-4o (via ChatGPT) is free but limited. Which one is more cost-effective depends on your needs and budget.
• Weaknesses: To be honest, based on the input information and past experience, Midjourney V7 may still have a bit of a “personality” in handling text generation (adding accurate text to images) and very precise finger details. Sometimes there will be some strange text or unnatural limbs. This is also a problem that AI drawing is still trying to overcome.

In short, there is no “best” one, only the one that “best suits” your needs. If you are pursuing top-notch artistic sense and stylization, you may choose V7; if you need accurate understanding of instructions or a free trial, you can check out GPT-4o; if you need convenient modification functions, maybe Gemini can help.

So, is V7 worth looking forward to?

The answer is: absolutely worth it!

The Alpha version of Midjourney V7 has shown amazing potential. Whether it is the surprisingly efficient draft mode, the significantly improved image quality, or the more understanding personalized functions, it makes people full of imagination for the future of AI drawing.

Although it is still in Alpha testing, some functions are not yet in place, and there may even be many bugs. But it is foreseeable that with continuous updates and improvements in the coming weeks, V7 will become more powerful and stable.

The combination of artificial intelligence and human creativity is developing at an unprecedented speed. Midjourney V7 is undoubtedly an important milestone in this wonderful journey. Let us wait and see what other surprises it will bring!