DALL·E
By Ryan Khalil · OpenAI (in ChatGPT) · Images right inside your chat. · Reviewed June 2026
DALL·E is OpenAI's image model, built into ChatGPT. Its superpower is convenience — you describe an image while you chat, see it, and refine it in plain conversation. Great for quick visuals and decent at putting readable text in images.
What is DALL·E good at?
Fast, convenient images without leaving your chat. Iterating by just talking ('make it brighter', 'add a dog'). Rendering short bits of text in an image more reliably than many tools. Good for social graphics, mockups, and quick concepts.
How do I prompt DALL·E well?
- Describe subject, style, mood, and setting in one clear sentence.
- Refine conversationally instead of rewriting from scratch.
- Name a style ('flat illustration', 'watercolor', 'photorealistic').
- Ask for a specific aspect ratio or use ('a wide banner', 'a square logo concept').
What are common mistakes when using DALL·E?
Expecting Midjourney-level artistic polish — DALL·E trades some beauty for convenience and control. For fine art and mood, use Midjourney; for quick and integrated, DALL·E wins.
How do I get started with DALL·E?
Try this in ChatGPT: 'Create a clean, flat-illustration banner of a lock turning into a key, purple and indigo, minimal.'