OpenAI started the new era in creating visuals with GPT-4O!

OpenAI announced the new visual creation features of the GPT-4O model. According to the company’s statement, the GPT-4O is capable of producing much more sensitive, detailed and realistic visuals than previous models. With this feature, users will only be able to create, edit the images they want with simple commands, or develop new designs through existing visuals.

The new era in creating visuals with GPT-4O!

OpenAI has long been advocating that the ability to create visuals should be a basic skill for language models. GPT-4O is also developed in line with this thought, the most advanced and useful visual creation system of the company. The visuals created with GPT-4O have become very useful not only in aesthetically, but also in terms of information transfer.

The new model understands the commands given by users more accurately and applies more precisely to the visuals. Especially in complex and multi-object visuals, GPT-4O performs better than its competitors.

For example, the model can now bring 10 to 20 different objects together in a single visual. In addition, it is much easier to produce informative images such as logos, diagrams, infographics, thanks to the ability to correct the texts and symbols of the visuals.

Examples shared by OpenAI include meeting grades on the white board, comic books, detailed infographics of scientific experiments, and visuals supported by meaningful texts. In the company’s statement, it was emphasized that creating visuals should be used not only for decorative purposes but also as a powerful tool in sharing information and communication.

The new GPT-4O model also has a multi-step visual production feature. In this way, users can develop the visuals they create with the model through a natural conversation. For example, the design of a game character can be shaped step by step and the consistency of the character can be maintained at each step.

The GPT-4O also has the ability to derive new images from these images by analyzing the images uploaded by the user. This feature makes the model a more intuitive and personalized tool for users. According to OpenAI, the variety and style of the images used in the GPT-4O model allows the model to create photo-real visuals and to convincing visual transformations.

OpenAI acknowledges that the new model has some limitations. There are some limitations in visuals containing graphics or multiple languages ​​containing very intense information, especially with small -sized texts. It was also stated that sometimes problems such as unwanted visual crops and inconsistencies may be experienced. The company said that improvements will be made in the future.

OpenAI also announced that it has taken various measures to secure the visual creation feature. All images produced by GPT-4O are added to C2Pa Meta data, indicating that the source of the content is OpenAI. In this way, the originality of the content created can be more easily verified. It was also emphasized that harmful content demands were automatically prevented.

As of today, the GPT-4O’s visual creation features have been presented as the default option for Plus, Pro, Team and free users in Chatgpt. Enterprise and EDU users will also benefit from this feature in a short time.

For DALL · E lovers, this model can still be used through a special Dall · E GPT. In addition, developers will be able to use GPT-4O’s visual creation feature through API in the coming weeks.

You can click here for all the details.