DALL·E was OpenAI’s text-to-image model family. DALL·E 2 and especially DALL·E 3 made it possible for anyone to create detailed images from a sentence, and DALL·E 3 was built directly into ChatGPT.
OpenAI announced in November 2025 that the dall-e-2 and dall-e-3 API models would be removed on May 12, 2026, and recommends its GPT Image models instead. In ChatGPT, image creation now uses GPT Image, so the experience people associate with DALL·E continues under a new name.
Key features
- Conversational image creation in ChatGPT: describe an image, then refine it by chatting.
- Image editing to change parts of an uploaded or generated image.
- Accurate text rendering for posters, labels and diagrams.
- Style and aspect control through plain-language instructions.
- API access with gpt-image-2 for developers.
- Content provenance with C2PA metadata on generated images.
DALL·E pricing
| Where | Price | Image access | Notes |
|---|---|---|---|
| ChatGPT Free | $0 | Limited image creation | Uses GPT Image |
| ChatGPT Go | $8/month (US) | More image creation | May include ads |
| ChatGPT Plus | $20/month | Expanded image creation | Plus Thinking, agent mode and more |
| ChatGPT Pro | From $100/month | Unlimited image creation | Highest limits |
| OpenAI API | Pay as you go | gpt-image-2 | Billed per image and text token |
DALL·E 2 and DALL·E 3 are no longer available in the API. Developers should migrate to gpt-image-2, gpt-image-1 or gpt-image-1-mini.
What it is like to use today
Creating images in ChatGPT is the most beginner-friendly option available: no prompt tricks, just describe what you want and ask for changes. It is particularly good at images that need readable text, like social graphics and simple infographics.
For highly artistic or photographic results, Midjourney still leads, and Stable Diffusion offers more control if you run it yourself.
Who should use DALL·E?
If you already use ChatGPT, its built-in image creation is the easiest way to get images for posts, presentations and mockups. Developers who used the DALL·E API should switch to GPT Image models.
DALL·E vs alternatives
Midjourney is the top choice for artistic quality. Adobe Firefly is built for commercially safe work. Stable Diffusion is free and open. Canva combines image generation with full design tools.
See all AI Image Generation tools.
Related guides
Frequently asked questions
Is DALL·E still available?
The DALL·E 2 and DALL·E 3 API models were retired on May 12, 2026. Image generation in ChatGPT now uses OpenAI’s GPT Image models.
Is DALL·E free?
You can create images free in ChatGPT with limits. Paid ChatGPT plans offer more image creation.
What replaced DALL·E 3?
OpenAI recommends gpt-image-2, gpt-image-1 or gpt-image-1-mini in the API, and ChatGPT uses GPT Image.
How do I make images with ChatGPT?
Ask ChatGPT to create an image and describe it. You can then ask for edits in the same conversation.
Is DALL·E better than Midjourney?
GPT Image in ChatGPT is easier and better with text in images, while Midjourney usually produces more artistic results.
