Creating images with ChatGPT
Summary
ChatGPT can generate original images from text descriptions, allowing users to quickly create and iterate on visual concepts. To get good results, write clear prompts (1-3 sentences) that specify the image's purpose, main subject, setting, and visual style, using direct language like 'soft natural light from the left' rather than vague phrases. The best way to improve images is through small, targeted revisions focusing on one element at a time, with clear spatial language and specific instructions for text or layout details.
Classification
Affected Vendors
Related Issues
Original source: https://openai.com/academy/image-generation
First tracked: April 10, 2026 at 02:00 PM
Classified by LLM (prompt v3) · confidence: 85%