OpenAI Launches GPT-4o Image Generation with Enhanced Text and Instructions

Abdullah Mustapha

March 26, 2025

OpenAI’s GPT-4o, released a year ago, has seen major improvements. One of its most exciting features is image generation. This allows the AI to create detailed images based on simple text instructions. Unlike older models, GPT-4o lets users refine their images step by step until they achieve the perfect result.

OpenAI’s GPT-4o: Smarter AI-Generated Images

Older AI models struggled with text in images. If you asked for a sign, you’d often get nonsense scribbles instead of readable words. GPT-4o improves on this. It understands instructions better and generates clearer, more accurate images.

Most AI image tools work by generating an image from a prompt, then tweaking the prompt to refine it. GPT-4o works differently. You ask for an image, then adjust it by simply giving new instructions. For example, you could ask for a sunset, then tell the AI to make the sky brighter or add birds. This step-by-step editing makes the process easier and more interactive.

Easy Customization

GPT-4o can create images from scratch, but it also edits existing ones. If you upload a picture of a cat and ask for a detective hat and monocle, the AI will add them seamlessly. You can keep refining the image—adjusting lighting, adding effects, or transforming it into a unique style.

Join GizChina on Telegram

Another great feature is image combination. The AI can merge elements from multiple images into a single composition. OpenAI claims it can handle 10-20 objects in a scene without issues. Most other AI models struggle with more than 5-8 objects.

Recognizing Weaknesses

Despite its strengths, GPT-4o isn’t perfect. OpenAI admits that it sometimes crops images incorrectly. It also struggles with generating very complex scenes or working with non-Latin text. Occasionally, the AI creates unrealistic or nonsensical details, known as hallucinations.

The Future of AI-Generated Art

AI image tools are becoming more powerful and accessible. GPT-4o is a big leap forward, offering an easy way to create and modify images. Whether you’re an artist, designer, or just experimenting, it’s a great tool for bringing ideas to life.

Read Also: Nintendo Switch 2 new patent suggests an AI-powered image enhancement is coming

To see GPT-4o’s image generation in action, OpenAI has shared video demonstrations. These show its capabilities and highlight how AI is evolving in the creative world.

Disclaimer: We may be compensated by some of the companies whose products we talk about, but our articles and reviews are always our honest opinions. For more details, you can check out our editorial guidelines and learn about how we use affiliate links.

Source/VIA :

Openai, Gsmarena