Product Updates
See what changed in GPT Image 2, including new AI image and video prompt workflows, product improvements, and fixes.
v1.8.2
Image and Video Reverse Prompting
New featureImage to PromptVideo to PromptReverse Prompt
What's new
- Added Image to Prompt for turning reference images into structured GPT Image 2 prompts.
- Added Video to Prompt for extracting scene, camera, lighting, composition, and motion details from reference videos.
- Prompts now separate subject, visual style, environment, camera, lighting, motion, and constraints so they are easier to copy and edit.
- Results include reusable prompt formats for image generation, video generation, creative remixing, and prompt learning.
Why it matters
Reverse prompting helps creators move from inspiration to an editable prompt faster. Instead of guessing how to describe a reference image or video, you can upload the media, review the structured prompt, then refine it for your own product, character, scene, or campaign.
Try it
Use Image to Prompt when the reference is a single image, product photo, poster, UI screen, or style frame. Use Video to Prompt when motion, camera movement, shot structure, or pacing matters.