The AI image generation model is better than ChatGPT.

The V7, MidJourney's model, is said to be capable of producing details like hands more consistently. Photo: Xeophon .

Midjourney, one of the first AI-powered image creation tools on the web, has just launched its first new model in almost a year.

The model, named V7, was released in beta testing at midnight on April 4th (US time). A week earlier, OpenAI had also launched a similar feature with ChatGPT, which quickly became a social media sensation, creating a new trend of creating Ghibli-style photos.

If you're new to Midjourney, you'll first need to evaluate the 200 images the tool suggests, based on your preferences. Then, the tool will create a personalized profile to produce images that match your aesthetic. V7 is the first Midjourney model to have personalization enabled by default.

You can enable or disable V7 by clicking on the “version” toolbar on the website. Additionally, this model can be used in Midjourney's Discord server.

Midjourney CEO David Holz described V7 as a “completely different architecture” in a post on X. “V7 is much smarter when it comes to handling text statements,” Holz further announced on Discord.

He commented that the model also handles original images quite well. Image quality is noticeably improved with a high degree of detail. From body parts like hands and feet to freckles and ceramic patterns, everything is consistent down to the smallest detail.

Some sample images were created using the V7 model. Photo: X/MidJourney.

In addition to the two modes: Turbo (more expensive) and Relax, V7 supports a new tool called Draft Mode, which allows image creation up to 10 times faster and at half the cost of standard mode. Images in Draft mode will have lower quality, but users can upscale and recreate images with just one click.

According to Holz, some standard Midjourney features are currently unavailable with the V7, including upscaling and retexturing. He stated that these features will be added in the near future, possibly within the next two months.

“This is a completely new model with its own strengths, and perhaps a few weaknesses as well,” Holz wrote on Discord. He noted that V7 might require a new command-line writing style and encouraged everyone to try the version out to help provide constructive feedback.

Techcrunch reporter Kyle Wiggers commented that the V7 could execute his basic commands well. “Midjourney’s model isn’t specifically optimized for the Ghibli style, at least not officially, but it can still produce aesthetically pleasing works,” he added.

According to AI experts in Vietnam, ChatGPT's image generation feature may be easier and more convenient to use than earlier image generation tools like MidJourney and Stable Diffusion. However, OpenAI's model has not yet created a breakthrough in image quality and may be inferior to previous tools.

Midjourney is a rather unique company. Founded in 2022 by Holz, co-founder of PC peripheral manufacturer Leap Motion, Midjourney has never received any outside investment.

The company also doesn't disclose its business performance, but CBInsights estimated Midjourney's revenue at the end of 2023 to be around $200 million . Recently, the company stated it's forming a hardware team to work on several undisclosed projects and continuing to train previously announced models to create 3D videos and objects.

Like ChatGPT, MidJourney is facing several lawsuits alleging violations of the rights of millions of artists by training AI tools on images collected from the web without the consent of the image creators.

Source: https://znews.vn/mo-hinh-tao-anh-ai-tot-hon-chatgpt-post1543472.html