OpenAI’s new image generation in ChatGPT (available to Plus and Pro users since 25th March 2025) has opened up some genuinely useful creative and professional workflows. Beyond fun and novelty, it’s being used to mock up ideas, edit images, create social content, and generate custom visuals that save hours of manual design work.
Here’s a rundown of 10 things worth trying in the new image generation model:
1Cast Anyone in Any Movie
Give ChatGPT the movie poster you want to emulate and pictures of who you want to cast in them!
2Create an Ad for Anything
Simply feed ChatGPT examples of your favourite ads and then pictures or just written details of your product or service.
3Create a Comic Strip
Describe the style of comic strip you’d like, with some outline of the narrative and any text you want in the form of captions or speech bubbles.
4Infographics
Start by building the written brief for your infographic in a regular text conversation with ChatGPT. Once you’re happy with that, ask it to produce an infographic in any style you like, even in your brand, based on the information.
5Put a Screenshot or Mockup onto a Screen
Need a lifestyle shot of someone using your app or website? No problem! Combining images is something ChatGPT’s new image generation model is pretty good at, so just by giving it your screenshot, and a reference image of someone with a device, you should get the shot you need.
6Product Shots
Whilst not always 100% perfect, the team behind this new AI image generator have clearly put a lot of time into making the model recognise objects and characters, so it can re-build them in other contexts e.g. in different settings, styles and from different angles or in different lighting.
7Complex Text Images
Given how bad AI has generally been with text in images, this is a huge leap for ChatGPT’s new image generation model. It can create pretty complex images containing text, far beyond what we’re used to seeing in Midjourney and the like. Yes, Ideogram have been leaning into doing text well for a while, but it can’t match some of the stuff we’re seeing come out of ChatGPT.
8Book Covers
Want to visualise that book you’ve been meaning to write for years? Maybe this is the trigger to get it done – to see it in the flesh (well, paper). Users are loving the new model’s capability, again largely with text, but also with tasteful design choices and mock-ups of physical printed materials like books and cards.
9App or Website Design
Bringing an idea to life quickly can be the difference between getting it built and not. The new model seems particularly good at understanding commands around interface design.
10Diagrams
Need a little diagram to explain something in a PowerPoint slide, or even just in an email to a colleague? Create one of these in seconds and watch their jaw hit the floor.
11Memes
And finally, of course, memes! RIP meme generator apps, because ChatGPT can easily handle turning any image and/or idea into a viral-friendly meme. The OpenAI team themselves pointed this out in the product demo when they launched, creating a meme of themselves.
Have a play at chat.openai.com – just make sure you hit the 3 dots and that you can see you have access to the new image generation model.
