OpenAItbh has launched an integrated image generation feature within ChatGPT, allowing users to create images directly using GPT-4o. This marks a significant evolution in AI-driven creativity, removing the need for users to rely solely on DALL·E or external platforms for image generation.
The feature, rolled out on March 25, is now available across ChatGPT’s Free, Plus, Pro, and Team subscription tiers. OpenAI described it as a major step toward making image creation a seamless part of AI-powered communication.
OpenAI’s CEO, Sam Altman, highlighted the breakthrough on X, calling it “an incredible technology/product.” He recalled seeing the first images generated by GPT-4o and being amazed that they were entirely AI-created.
“We think people will love it, and we are excited to see the resulting creativity,” Altman stated. He acknowledged that while the technology enables remarkable content creation, it also brings challenges, as some outputs may offend people. However, he emphasized OpenAI’s commitment to giving users more control over their creative freedom while monitoring societal expectations.
Unlike previous AI models, GPT-4o’s image generation is more precise, interactive, and adaptable. OpenAI highlighted several key improvements:
– Text Rendering: AI-generated images now include clearly legible text, making them ideal for infographics, diagrams, and labeled visuals.
– Multi-turn Generation: Users can refine images through conversation, modifying details while keeping consistency across versions. This is particularly useful for tasks like storyboarding, branding, and character design.
– Instruction Following: GPT-4o processes complex prompts with greater accuracy, allowing for detailed compositions with up to 20 distinct objects while maintaining correct relationships between them.
– In-Context Learning: The model can analyze uploaded images and use that context to create new visuals, making it valuable for design inspiration and brainstorming.
– Knowledge Integration: GPT-4o combines text and image understanding, enabling it to generate visuals suited for technical diagrams, educational illustrations, and other context-specific imagery.
Developers will soon gain API access to GPT-4o’s image generation capabilities, enabling broader integration across applications. While the feature is already available for ChatGPT users, OpenAI plans to expand access to Enterprise and Education subscribers.
Users can generate images by simply describing them in ChatGPT, specifying details such as colors, aspect ratios, and design preferences. However, OpenAI noted that due to the complexity of GPT-4o, rendering images could take up to a minute.
For businesses and developers needing more customization, OpenAI confirmed that DALL·E will remain available as a separate model option.
With this advancement, OpenAI is reinforcing its position at the forefront of AI-generated content, offering users more creative control while pushing the boundaries of what’s possible with artificial intelligence.
