The new MidJourney V6 is crazy and the images of this AI are cinematic: this is how it compares to its previous version
The world of artificial intelligence ends the year in style with the recent landing of the V6 version of Midjourney, which, just by seeing some images generated with this tool, You can see the astronomical leap it has made compared to its previous version, V5.2, which came out in June of this same year. —difference of only six months—.
Since its first version was released in March 2022, this generative tool has slowly demonstrated its power and, with V6, the potential of Midjourney to create images that look like real photographs is clear.
Now it has brought a series of enormous improvements to the table with the new versionas shared in an official Discord thread, reserved for members of the exclusive MidJourney club.
Notable updates include more precise build, longer prompts, improvements to model consistency and awareness, improved build remixes, and the addition of text to images, which is worth highlighting because it’s been a big issue in the past. .
This is how this new version of Midjourney compares to the one from six months ago: like a movie
The first aspect worth mentioning is the new model’s ability to follow detailed instructions. Compared to V5.2, V6 shows a great improvement in the coherence and organization of elements in the generated image.
Although both models have their successes and failures, but the V6 stands out above and beyond for its ability to interpret and capture the instructions you give it to the minimum details..
For example, since HowtoGeek provided the following text: “Draw a market in a futuristic city. To the left of the painting is a woman with a shopping basket in her left arm. To the right is a flea market stall. Behind the fruit stall is “a robot. The robot is purple and holds an apple in its outstretched right arm.” This was the result given by V 5.2 and V6.
HowtoGeek
On the other hand, as mentioned before, the ability to integrate text into images is a feature that certainly marks a big step forward, even if it seems silly. MidJourney V6 solves this big problem, allowing users to include text in their instructions.
Although it is not yet perfect, as they explain from HowtoGeekwith a 75% success rate on the first attempt, represents a clear advance compared to V5.2, which strives to obtain precise results in this aspect, but finds it difficult to achieve it.
Regarding artistic quality, it is again essential to consider how the V6 revolutionizes the scene and stands out in this aspect. The evolution is evident in the composition and details, almost unrecognizable to the human eye as something generated with artificial intelligence.
Although it is still in an alpha phase, Just by looking at the following image you can perceive a great wealth of details and a realism that leaves you with your mouth open..
Nevertheless, It must be remembered that the tool is expected to improve even more. As David Holz, Midjourney’s CEO, noted in his Discord message announcing the new model “it’s an alpha test. Things will change frequently and without warning… They will change significantly as we get to the full release of V6. “V6 is not the final step, but we hope everyone feels the progression of something profound that is deeply intertwined with the powers of our collective imagination.”
How can I access MidJourney V6?
As with previous versions of MidJourney, the sixth version is available through Discord. Once you have completed the subscription payment, you will gain access to the tool, where you can unleash your imagination.
To access V6, simply select it as the default model in settings or add “–v 6” to the end of any message. This will ensure that you use the correct model.
