# Can ChatGPT Create Images?

> Yes, ChatGPT creates and edits images with OpenAI’s built-in image model, and how many you can make depends on your ChatGPT plan. You describe the picture in the chat, or upload a photo and say what to change. Kenerate's chat works the same way but lets you pick the model: 1–4 images per request from 6 models.

URL: https://kenerateai.com/can-chatgpt-create-images
Publisher: Kenerate AI (https://kenerateai.com)
Last updated: 6 October 2026

## Key facts
- ChatGPT: Creates and edits images with OpenAI’s built-in image model; limits depend on the plan
- How you ask ChatGPT: Describe it in a conversation, or More > Images; upload a photo to edit it
- Kenerate chat models: Seedream 5.0 Pro (default) · GPT Image 2.5 Sunburst · Nano Banana 2 · Qwen Image 3.0 · Flux 2 Pro · Ideogram v4
- Images per request: 1 to 4
- Reference images: Up to 10 per request (fewer on some models)
- Edits: Attach a photo and describe the change
- Remixes: 3 variations · Animate · Upscale 4K · 9:16 · 16:9 · Remove background · Layers · Camera angle · Expand · AI filter · Edit
- Control: Ask first (approval card with editable model, ratio and prompt) or Auto within a per-chat budget
- Where results go: In the chat and in your Image Studio history

## Steps
1. Open the chat — Go to Kenerate Agent at /app/chat and sign in.
2. Describe the image — Say what you want, how many (1–4) and the ratio. Attach up to 8 photos to edit or use as references.
3. Check the card — In Ask first mode an approval card shows the model, ratio and prompt; change any of them, then run it.
4. Remix or keep going — Tap a remix such as 3 variations, Animate or Upscale 4K, or ask for the next change in plain words.

## In depth

### How ChatGPT makes images
ChatGPT's image generation is built into the chat. You ask for an image in plain words, or open More and choose Images, and OpenAI’s built-in image model draws it. You can upload an image and describe changes, start from a template, and on mobile turn a sketch into a picture. OpenAI's help centre lists the feature on every ChatGPT tier, on the web, iOS and Android.
Because the model is chosen for you, ChatGPT is simple: one place, one image model, and the rest of your conversation right there.

### How Kenerate's chat makes images
Kenerate Agent calls the same image engine as Image Studio. Each request makes 1 to 4 images with the model you name, or Seedream 5.0 Pro by default, and can use up to 10 reference images. In Ask first mode you see an approval card before anything runs and can change the model, ratio or prompt; in Auto mode it runs within the budget you set for that chat. Every result is also saved to your Image Studio history.

### Picking a model in the chat
Name the model in your message, or change it on the card.
- Seedream 5.0 Pro: photoreal people, products and places, up to 4K.
- GPT Image 2.5 Sunburst: layouts, posters and pictures that need readable text.
- Nano Banana 2: the same face or product across a set of images.
- Qwen Image 3.0: quick drafts, with English and Chinese prompts.
- Flux 2 Pro: artistic, clean compositions.
- Ideogram v4: typography, logos and graphic design.

### After the first image
Every image in Kenerate's chat has one-tap remixes. 3 variations, Animate, Upscale 4K, Vertical 9:16, Wide 16:9, Remove background and Split into layers send straight away; New camera angle, Expand, AI filter and Edit fill in the message so you can finish the instruction.

### ChatGPT or Kenerate's chat: which fits
Both make images from a conversation. They suit different people.
- ChatGPT: Best when you already write, plan and research in ChatGPT and want a picture in the same thread, with one built-in image model chosen for you.
- Kenerate AI chat: Best when you want to choose the image model for each job, use up to 10 references, and carry on into video, upscaling or a deck in the same chat.
- ChatGPT photo editor guide: Best if your question is really about editing a photo: how ChatGPT does it and where a brush edit works better.
- Image Studio: Best when you'd rather set the model, size and references yourself without a chat in between.

References:
[1] OpenAI Help Center: Images in ChatGPT (checked 6 Oct 2026): https://help.openai.com/en/articles/11084440-images-in-chatgpt

## FAQ
Q: Does ChatGPT generate images?
A: Yes. ChatGPT generates images with its built-in model, OpenAI’s built-in image model. Ask for a picture in any conversation, or open More and choose Images, and it creates what you describe. OpenAI's help centre says ChatGPT Images is available on every ChatGPT tier, on the web, iOS and Android.

Q: Can ChatGPT edit my photos?
A: Yes. Upload a photo and describe the change, and ChatGPT makes an edited version. Our ChatGPT photo editor guide covers that in detail, including area selection and the brush tools on Kenerate if you need precise edits.

Q: Is image generation in ChatGPT free?
A: Image generation is available in ChatGPT, but how many images you can make depends on your plan and OpenAI changes those limits from time to time. Check OpenAI's own pages for the current numbers.

Q: Which AI model does ChatGPT use for images?
A: ChatGPT uses OpenAI’s built-in image model, OpenAI's built-in image model. In ChatGPT you don't pick between image models. In Kenerate's chat you can: Seedream 5.0 Pro is the default, and GPT Image 2.5 Sunburst, Nano Banana 2, Qwen Image 3.0, Flux 2 Pro and Ideogram v4 are one word away.

Q: How do I make images in Kenerate's chat?
A: Open /app/chat, describe the picture and say how many you want (1 to 4) and the ratio. You can name a model in your message or change it on the approval card. The images appear in the chat and are also saved to your Image Studio history.

Q: Can I use my own photos as references?
A: Yes. Attach up to 8 files to a message and the chat can use up to 10 reference images in one request, for a product, a face or a style. Some models take fewer: Qwen Image 3.0 and Flux 2 Pro use up to 3, and Ideogram v4 works from text only.

Q: What can I do with an image after it's made?
A: Every image has one-tap remixes: 3 variations, Animate into a short video, Upscale to 4K, recreate in 9:16 or 16:9, Remove background, Split into layers, a new camera angle, Expand, an AI filter, or Edit. Each one sends a new request in the same chat.

Q: Can the chat make other things besides images?
A: Yes. The same chat can make video, music, voice-overs, talking avatars and 3D models, and write PDF or DOCX documents, PowerPoint decks and spreadsheets. Video and 3D need a paid plan.

## Related pages
- ChatGPT photo editor: https://kenerateai.com/chatgpt-photo-editor
- ChatGPT alternatives: https://kenerateai.com/chatgpt-alternatives
- Kenerate Agent: https://kenerateai.com/kenerate-agent
- AI image generator: https://kenerateai.com/ai-image-generator
- GPT Image 2 vs Nano Banana: https://kenerateai.com/gpt-image-2-vs-nano-banana

## Sources
- OpenAI Help Center: Images in ChatGPT (checked 6 Oct 2026): https://help.openai.com/en/articles/11084440-images-in-chatgpt