ChatGPT helped popularize generative AI by making advanced text generation widely accessible.

Now, it’s expanding its creative range with improved image generation capabilities. From producing visuals in trending style — like Studio Ghibli — to making quick photo edits or generating fictional web pages from a simple text prompt, ChatGPT continues to show its versatility.

With these new tools in mind, let’s explore how to create images using ChatGPT, what it does well, where it still has limitations, and practical ways to work around those gaps.

*To generate images with ChatGPT, you’ll likely need a paid subscription (starting at $20 per month). Free users can still access image generation through the GPT-4 model, but they face stricter limits and lack access to certain features, including file uploads, advanced editing tools, and priority access during peak times.GPT-4

Table of Contents

• How to Use the ChatGPT Image Generator
• What Type of Images Can You Create Using ChatGPT?
• Strengths and Weaknesses
• Frequently Asked Questions

How to Use the ChatGPT Image Generator

Generating an image with ChatGPT is simple, but there are a few things to keep in mind for the best results.

The most important factor is the specificity of your prompt.

Like other AI image generators, ChatGPT cannot think for itself or make assumptions, so it's up to you to provide clear and intentional instructions if you want a particular outcome.

Generating an image with ChatGPT is straightforward, but getting the best results depends on how you craft your prompt. The key is specificity.

Like other AI image generators, ChatGPT cannot think for itself or fill in creative gaps — it relies entirely on the instructions you give.

The clearer and intentional your prompt, the more likely you are to get the outcome you want.

1. Create a New Chat

To start generating an image with ChatGPT, open a new chat. This happens automatically when you launch ChatGPT, but you can also start a new chat manually by selecting the pencil icon at the top of your screen.

Using a new chat helps keep your image generations separate from your regular conversations. While it’s possible to generate images in the middle of a longer conversation, it becomes harder to scroll back and find them later.

2. Write a Detailed Prompt

In the prompt box, describe the image you want.

It is important to include a phrase like “Generate an image” to signal that you’re requesting a visual output. Keep in mind, even small wording changes can have large effects on your results.

For example, changing the phrase to “Create a drawing” or “Paint a picture” can completely alter the style, even when the rest of the prompt stays the same.

Below is a comparison using the same base prompt:

“A cat swinging a baseball bat, narrowly missing a baseball.”

Each version uses a different lead-in phrase:

1. Create a drawing
2. Paint a picture
3. Generate a photo
4. Create a graffiti image

As you can see, small details like this make a big impact on your final generation. Intentional, descriptive language is key to generating a successful image.

When writing prompts of your own, use this simple formula:

• Describe your subject +
• Describe what they’re doing and any surrounding details +
• Specify the style of image

3. Wait for the Image to Generate

Image generations typically take under a minute, though more detailed prompts may take longer. Occasionally, a generation will fail and prompt you to try again. In most cases, repeating the request works without issue.

However, if you see a message that says “Getting started” and nothing happens for more than five minutes, it may be stuck. If refreshing doesn’t resolve the issue, you’ll need to open a new chat to continue. While this bug is hopefully rectified soon, it currently disables all chat features within the conversation, requiring you to begin a new conversation.

4. Edit or Save Your Image

Once your image is ready, you can either save it or make edits.

To save the image:

• On desktop, hover over the image and click the download icon in the top right corner.
• On mobile, tap and hold the image until your device gives you the option to save.

If the image isn’t quite what you had in mind, you can request an edit. On desktop, clicking the image opens a dedicated editing window, where you can describe the changes you’d like to see. There’s no need to retype your original prompt — just explain the revision, and ChatGPT will adjust the image accordingly.

Below is an example of this. After producing an original image of a boy playing soccer, I asked ChatGPT's image generator to: “Change the background to include some trees.”

Naturally, ChatGPT will understand that I am referring to the previously created image.

The new image will be generated in the original conversation thread.

Here you can see the difference between the two images.

It’s important to place your editing prompt directly after the image you want to update. If there are other messages or generations in between, ChatGPT may have trouble identifying which image you’re referring to.

As with your original prompt, specificity is key when requesting edits. Rather than giving vague instructions like “make it darker,” try something more descriptive — for example, “increase the shadows in the image and slightly reduce the overall brightness.”

You can also ask ChatGPT to remove specific objects from an image. As long as you clearly identify what you want removed, the tool is typically very effective at making accurate adjustments.

What Type of Images Can You Create Using ChatGPT?

ChatGPT's image generator is capable of producing a wide range of image styles, from realistic portraits to stylized artwork.

Whether you're experimenting with creative prompts or building assets for personal use, here are some specific areas where ChatGPT performs well, along with prompt ideas to help you get started.

Portraits, Headshots, Profile Pictures, and Characters

ChatGPT can generate images of real or fictional people in a variety of styles, including photorealism, cartoon, digital painting, and professional headshots.

Creating a headshot with ChatGPT is an accessible way to quickly produce an image for use on resumes, portfolios, or business profiles.

To do this, upload a clear photo of yourself and ask ChatGPT to:

"Generate a professional headshot using your image as a reference."

The result will be a clean, portrait-style image that is often ready to use with little to no adjustment.

As with any generated image, you can request edits such as changing clothing, adjusting facial expressions, or modifying the background.

Keep in mind that these image generations can be inconsistent and may take a few tries to produce a convincing result. If a generation turns out poorly, it's usually better to request a new one rather than attempt edits.

Static Advertisements

Another recent advancement in ChatGPT image generation is its ability to create cohesive images that include both graphics and text. This is a major improvement in AI image creation, since earlier models often struggled with text accuracy.

Now, digital marketers and small business owners can use ChatGPT to create mock advertisements, product promos, or event banners. These images work well for social media previews, pitch decks, or quick test campaigns.

To create a compelling image, upload a photo of your product and include your brand name in the prompt. ChatGPT will use both inputs to generate a customized, branded result.

Social Media Graphics

ChatGPT can also generate visual assets for social media posts, including Instagram stories, YouTube thumbnails, and Pinterest pins.

While it does not support exact aspect ratios, you can still request general formatting like vertical, square, or horizontal layouts.

For the most personal results, try prompting ChatGPT to create a graphic while leaving space for you to add your own images. For example, I generated this Instagram Valentine's Day template with the prompt:

"Create an Instagram post for Valentine's Day with a bold pink and red theme. Include two empty photo frames or placeholders where personal images can be added later. Add the text 'Happy Valentine’s Day' in a large, elegant font at the top."

The result is a ChatGPT-generated image that can be customized using a free online image editor, where you can drop your own images into the spaces.

It's the perfect mix of AI-generated images and your own creativity.

Also, remember that images created with ChatGPT do not have transparent layers, so you may need to use a transparent background maker in combination with the tool.

Stylized Artwork and AI Memes

AI-generated artwork has become a core part of modern meme trends.

ChatGPT works well for creating stylized visuals that match formats like the Studio Ghibli trend or the Starter Pack Action Figure meme.

These are usually created by pairing a descriptive prompt with a reference image or theme.

You can also use ChatGPT to combine existing images into a new, cohesive composition.

For example, you can upload a background image and a subject, then ask ChatGPT to blend them into a single scene.

Strengths and Weaknesses

ChatGPT's image generator is surprisingly capable for a text-based tool, especially when you need quick, stylized visuals or want to experiment with trending design areas. However, like any AI tool, it has some limitations to be aware of, particularly when it comes to precision or consistency.

Strengths

• Fast Image Generation: Most images are generated in under a minute. Even when an image is taking a while, you can continue working in another conversation while it generates in the background.
• Strong Prompt Interpretation: While the level of specificity required might be intimidating to some, it is also reassuring how well ChatGPT can follow complex prompts to incorporate all requested elements.
• Capable of Generating Any Image Style: You can create visuals in nearly any style, including realism, cartoon, cyberpunk, pixel art, graffiti (my personal favorite), or watercolor. Just include your desired style in the prompt, and ChatGPT will do its best to match it.
• Built-in Editing Support: Once an image is generated, you can easily ask for edits without having to download and upload files. This is great for making small changes within an image that would otherwise require complex edits.
• Great for Trending Content and Memes: If you're creating a standalone image or jumping on a current meme trend, ChatGPT can generate quick, one-off visuals to help you join the conversation fast.

Weaknesses

• No Aspect Ratio or Resolution Control: While you can specify that you want an image to be landscape or vertical, ChatGPT cannot create images with specific dimensions. As a result, we recommend using an online image resizer to adjust final images for best display when posting on social media.
• Inconsistent Edits: Sometimes, you will ask ChatGPT to add text to an image, but it will instead make other unrequested edits. In these cases, you simply need to make a new request, which takes up time and sometimes does not work (it can be frustrating).
• No Layer Editing: Each image is generated as one layer. As such, you cannot change individual details like resizing a subject or changing a font. This is especially problematic when you're trying to repurpose an image for different applications.
• No Export Options: All images downloaded from ChatGPT are configured as a PNG file by default. There are no options to control compression or export images in other popular file types like JPEG.
• Images Only: ChatGPT currently only supports image generation and edits. This means that users cannot generate AI videos or edit GIFs without using a third-party tool.

Frequently Asked Questions