OpenAI has released an addendum to the GPT-4o System Card detailing the capabilities of GPT-4o image generation. This new model represents a significant leap forward from the previous DALL·E 3 series, offering significantly more capable and sophisticated output. The primary enhancement is the ability to create photorealistic images, bridging the gap between AI-generated content and real-world photography.

Furthermore, GPT-4o introduces advanced multimodal functionality; it is not limited to text-to-image prompts but can also accept existing images as inputs. This allows users to transform uploaded visuals, enabling complex editing, style transfer, and object manipulation within a single workflow. This evolution matters for the industry because it moves generative AI beyond simple creation tools into comprehensive visual editing suites.

By integrating photorealism and image transformation into a unified model, OpenAI is setting a new standard for AI utility in creative and professional sectors.