OpenAI goes one step further in the field of artificial intelligence with its new image production feature in the Chatgpt. This feature, called “4O Image Generation ,, is activated for all users as of today. Users will now be able to create and edit direct visuals via Chatgpt using the GPT-4O model.

The new feature is offered to chatgpt to both paid and free users. There will be a specific production limit for free users, but its limit will vary depending on the current demand and intensity. Previously, free users could produce three visuals a day through DALL-E 3.

As the GPT-4O is known, it has the capacity to produce various data types such as text, image, sound and video. This increases the details in the visuals. One of the remarkable improvements of this model stands out as binding binding ”(binding). This feature allows artificial intelligence to correctly understand the complex object and feature relationships. For example, while most image models can mix the colors and shapes given in a request, the new system can correctly connect the 15 to 20 objects. In addition, the model can learn from the uploaded visuals and use them as a reference.

Another powerful aspect is the text processing side. Traditional artificial intelligence models often create text in visuals, while making spelling errors, GPT-4O significantly reduces these errors. OpenAI said that he trained the GPT-4O on the registered data he obtained from his partnerships with companies such as “publicly open data” and Shutterstock to support the new image feature.

Autoregress approach in production

Most visual-producing models, such as DALL-E, use the diffusion model technique that creates the whole image at one time. However, OpenAI uses an authorized approach that goes to a difference and creates images line line and column columns. This technique difference increases accuracy, especially in complex text and object relationships.

The new feature can respond to complex visual demands such as scientific diagrams, very panel comic books and information posters. It can also be used for practical designs such as transparent background stickers, restaurant menus and logos. Therefore, the new vehicle appeals to both professional and personal use. In addition, OpenAI is also referred to the world information of the model during production. When you want Newton’s prism experiment, it can produce visuals without detail. If you want to make text descriptions in the visual. Visual production lasts a little longer than before.

Security in the foreground

OpenAI emphasizes that it takes comprehensive security measures to prevent the abuse of visual production tool. The system prevents the production of obscene content and rejects attempts to remove copyright signs.

Although there is no direct waterplace in the visuals, the OpenAI points out that they are produced by artificial intelligence by using C2Pa commodity data in all visuals.

4o Image Generation How to Use?

4O Image Generation is presented to Plus, Pro, Team and free users as the default image formed at Chatgpt. It may take some time for it to be activated in your account. Enterprise and Edu will soon be accessed. It can also be used in Sora.

Developers will soon be able to produce images with GPT – 4O via API and access will be provided within the next few weeks.

Creating images and customizing the chatgpt as easy as chatting. You can produce visuals by explaining what you need, including any details, such as exact colors or a transparent background using only the height ratio, HEX codes.

You can browse the sample images from the gallery below.