Ope opening Has announced it Chadds You can create realistic images directly from the chat, avoiding the use of external tools like doll-e. Based on the activity model GPT-4O This, according to the company, uses an autonomous approach and its true world knowledge Provide professional status results more intuitive and effective.
That is, it allows them to create Simple text descriptions In the conversation and Loaded filesIn addition to upgrading Representations of the textSymptoms and understanding of monitoring Environment. We remember that GPT-4O allows you to communicate with the text, audio and film, and it is expected to make it with the video.
Manages 10 or 20 different objects
Because the generation of the images is the native of GPT-4O, the sample conversation environment can be used for both images and text, thus The sync of the image is maintained at all times. This means that if you are designing a character, he is full and his appearance will be maintained in essence.
Says in Openai A Connection That is, there are difficulties with 5 to 8 objects for other organizations, GPT-4o can manage 10 or 20 different objects. They say that this is the fruit of a year of work with a hundred human trainers, and they can understand the model well, and allows you to create realistic images.
These are the main applications that include
- Image creation for work: Advertisement graphics for maps, info graphics, hexadesimal codes, logos, and complex methods.
- High text content images: Instruction posters, visualization of learning ideas, text brands, presentation cards.
- Results for external use with customization options: Personalized stock photos with an obvious background to slides.
- High quality photo pictures and realism: Advanced capacity for photosynthesis, including accuracy in lights, shades and settings.
- The ability to use an image as a starting point: A pet paint, portrait version, inspiration for interior decoration based on real space image.
- Images that use the conversation and the real environment: Bird posters found in Central Park, visualization of the era of art history discussed in the previous conversation.
Openai instant instant
The following example represents correctly The status of detail and accuracy When creating the image with the GPT-4O model.
Create a photosynthesis image of two twenty seconds witches (one of the gray palage and the other with a long and wavy red brown hair) who reads the traffic signal.
Environment: Any street of New York’s Williamburg, with a post covered with multiple comprehensive signals (for example, scanning tables, parking permits, trailer standards and trailer standards), including some absurd signals in the center: these signals are legal (paragraphing, thus legal).
The letters: A witch has a broom, and the other has a rolled magic carpet. They are in the front, the heads are slightly tilted to the camera’s back and posters.
Background Composition: Roads + Stopped Cars + Buildings -> Street Poster -> Sorcery. The characters should be as close as possible to the camera that takes the photo.
Availability
New generation pictures with GPT-4O will start processing for users Chadd Plus, Pro, Team Y Free The model 4 is also available as a predetermined image creation. Sora And positions Sadgift Enterprise y Ed They will soon have access to them.
Sign up with us Newsletter Get the latest messages on technology in your post.
