How to Use ChatGPT Images, Try Presets, and Generate Multiple Images at Once

OpenAI is doing something subtle but important with ChatGPT images. With the rollout of a new image model called GPT Image 1.5, the company hasn’t just improved output quality behind the scenes. It has also rethought how people actually use image generation inside ChatGPT.

Instead of hiding everything inside a regular chat, images now have their own dedicated space. After spending some time with it, this update feels less like a flashy upgrade and more like OpenAI finally smoothing out the rough edges of AI image creation.

Here’s how the new Images section works, what’s changed under the hood, and why it feels easier to use than before.

GPT Image 1.5 Is About Control, Not Just Quality

At the model level, GPT Image 1.5 focuses on three things users usually complain about: edits going off track, slow generation, and images that look fine at a glance but fall apart on closer inspection.

When you upload an image and ask for a specific change, ChatGPT is now better at editing only what you ask for. Lighting, background details, facial features, and overall composition tend to stay intact instead of getting unintentionally rewritten.

Text inside images has also improved, especially smaller or denser text that older models struggled with. OpenAI also says image generation is now significantly faster, which becomes obvious when you start testing multiple variations back to back.

The Images Section Changes How You Start

The bigger shift, though, is the new Images section itself.

On both mobile and web, ChatGPT now has a dedicated Images tab in the sidebar. When you open it, you’re not dropped into a blank conversation. You see a clean, visual-first interface.

At the top is a simple prompt bar that says “Describe a new image.” Below that is a row of style presets with clear visual examples. These include options like Bollywood poster, Festival, Navratri, Mithila, Jaipur textile, and Sari landscape.

Scroll a little further and there’s a “Discover something new” section with quick ideas like creating a cartoon, designing a holiday card, generating an album cover, or seeing what you’d look like as a K-pop star.

It immediately feels less intimidating, especially if you don’t enjoy writing long prompts from scratch.

How to Use ChatGPT Images on Mobile

If you’re on your phone, getting started takes seconds.

1. Open the ChatGPT app.
2. Tap the two-line menu icon in the top-left corner.
3. Tap Images below “New chat.”
4. Upload an image or describe a new one.
5. Choose a preset style or write your own prompt and submit.

That’s it. There’s no extra setup, and you don’t need to switch between chats to experiment.

How to Use ChatGPT Images on the Web

On the web, the flow is almost identical.

1. Open ChatGPT in your browser.
2. Click the sidebar icon in the top-left corner.
3. Select Images below “Search chats.”
4. Upload an image or enter a description.
5. Pick a style or add a custom prompt and generate.

One small but useful detail is multitasking. You can send multiple image requests at the same time. If one image is still processing, you can start another without waiting.

Presets Feel Like Built-In Prompt Banks

After playing around with the new Images section for a bit, this is where it really started to make sense for me.

I tried giving ChatGPT two images to process at the same time. One was a photo of myself, where I picked a fisheye-style preset just to see how aggressive it would be. The other was an image of an Apple Store, which I asked ChatGPT to turn into a post-apocalyptic version with everyone inside turned into zombies.

Both requests ran in parallel, and both stuck surprisingly close to what I asked for. The fisheye effect didn’t mess with lighting or facial features, and the Apple Store kept its structure while fully leaning into the apocalyptic theme.

That’s when it clicked. These presets aren’t just filters. They’re essentially prompt banks.

And that matters more than it sounds. When the Ghibli-style image trend took over earlier this year, half the effort wasn’t generating images. It was hunting down the “perfect” prompt, copying it from somewhere, tweaking it, saving it in notes, and repeating the whole process later.

That gets old quickly.

With the new Images section, you can start from a known style, see what it does, and then build on top of it. You’re no longer starting from zero every single time.

Source