# OpenAI Unveils Advanced Image Generation in GPT-4o

OpenAI has announced a major upgrade to GPT-4o's image generation capabilities, marking a significant leap beyond its previous DALL·E 3 models.

The new system can produce photorealistic images and accepts images as inputs for transformation, according to an addendum to the GPT-4o System Card posted by OpenAI on social media. This represents a fundamental shift in how the model handles visual content.

**What's Changed**

Unlike DALL·E 3, which focused primarily on text-to-image generation, GPT-4o's image generation can now work with existing images as starting points. This image-to-image capability allows users to modify, enhance, or completely transform uploaded photos and graphics.

The company emphasizes that the photorealistic quality represents a "significantly more capable" approach to image synthesis, suggesting improvements in detail,