OpenAI Introduces Native Image Generation in ChatGPT Using GPT-4o

The new feature allows users, including those on the free tier, to create and edit images directly in ChatGPT, with enterprise and API access coming soon.

OpenAI has integrated image generation into ChatGPT using its multimodal GPT-4o model, replacing the need for separate tools like DALL-E 3.
The feature, now live for all ChatGPT users, supports creating, editing, and refining images within conversations, leveraging user prompts and uploaded visuals.
GPT-4o excels in rendering text accurately within images, maintaining visual consistency, and producing photorealistic results, though generation times are slower than previous models.
OpenAI has implemented safeguards, such as metadata to identify AI-generated images and restrictions on harmful content, while addressing challenges like non-Latin script rendering and cropping issues.
Enterprise, education, and API users are expected to gain access to the image generation feature in the coming weeks, expanding its availability beyond current user tiers.