How to Use ChatGPT Images 2.5 (Step-by-Step Tutorial)

About this video

Kevin Stratvert tours ChatGPT Images 2.5, presenting it as faster and able to generate images from scratch, edit your own photos with precise changes, and create graphics with text without starting over. In Images he shows trending ideas and templates, using an ’80s flashback template that auto-writes a detailed prompt to preserve identity and facial features while changing hair, clothing, accessories and surroundings, then dragging in his own photo. He creates a Kevin Cookie Company ad from a text description alone, noting more direction gives more control. He then edits iteratively: using Comment to place a glass of milk exactly where tapped, Erase to remove a vase of flowers, and Markup with an arrow and box plus written instructions to add ‘The Kevin Cookie Company’ text wrapping naturally around a glass. Other tools include Remove background for a transparent background, Resize for a different aspect ratio, and a version history to return to earlier versions. He also demonstrates @sketch to turn a rough sketch into a photorealistic ad, transforming his photo into a professional baker’s outfit, combining it with a second reference image of a bakery monster so both appear together eating cookies, and making a poster reading ‘fresh from the oven’ and ‘buy two cookies, get one free.’ He warns text and small details can still be imperfect, but you can keep editing the same image to fix them.

Key takeaways

• Start in ChatGPT > Images from a template or a from-scratch description; templates show the detail level useful in your own prompts
• Use Comment to tap a spot and describe an addition, like a glass of milk placed exactly there
• Use Erase to remove unwanted objects and Markup to draw arrows/boxes combined with written instructions for precise text placement
• Use Remove background, Resize, and version history to change format without losing earlier versions
• Sketch with @sketch, upload photos to transform them, add a second reference image to combine subjects, and generate posters with text

Episode details: published September 16, 2026 · 9:45 long. Summary written by Self Growth Videos from the episode itself. Business results and any income figures mentioned are Kevin Stratvert’s own claims, shared for education — they are not typical or guaranteed.

← Previous: Claude Design Tutorial for Beginners · Next: How to Build AI Agents with Claude Code + UiPath for Coding Agents →

All Kevin Stratvert videos · Kevin Stratvert profile · Kevin Stratvert on YouTube