OpenAI is introducing image generation directly within ChatGPT. Powered by its flagship multimodal model, GPT-4o, the chatbot can now create visuals straight from the chat interface.
The feature will initially be available to ChatGPT Plus, Pro, Team, and free users. Enterprise and Education tier users will get access soon.
Today we have one of the most fun, cool things we have ever launched . . . native images in ChatGPT, OpenAI CEO Sam Altman said at the beginning of a video stream Tuesday. Altman acknowledged that the feature had been highly anticipatedespecially since competitors like Google Gemini have offered integrated image generation for some time.
ChatGPT now allows users to generate images based on prompts, conversations, and uploaded files. Users can create brand new images or transform existing images. OpenAI says the world knowledge trained into the GPT-4o model allows ChatGPT to better understand the contexts in which images are used. It is also better at following prompts rendering text within images, OpenAI says.
Users can refine images by prompting the model with natural language. For instance, when designing a video game character, the model can maintain visual consistency across multiple iterations as the user makes adjustments.
OpenAI says it expects people to use the tool work-related visuals that require precision (such as diagrams, infographics, branded content), text-heavy images (instruction posters, business cards), photorealistic images with accurate lighting and textures, and visuals that benefit from conversation context.
By simplifying the process with a single multimodal model that handles all image generation tasks, OpenAI is positioning ChatGPT as a go-to tool for both personal and professional image generation.
