If you have ever typed a sentence into ChatGPT and watched it turn into a finished image, you have used the technology that DALL-E became. DALL-E was OpenAI’s pioneering text-to-image model, but in 2026 the name on the box has changed: image generation inside ChatGPT is now powered by GPT Image, sold to users as ChatGPT Images. This review explains what DALL-E is today, what replaced it, what it costs, and whether it is still worth your time.
What Is DALL-E?
DALL-E is OpenAI’s text-to-image generation system. DALL-E 3, released in 2023, turned natural-language descriptions into images and lived inside ChatGPT, where it expanded short prompts into detailed visual instructions automatically. You did not need prompt-engineering skills: you described what you wanted in plain English and the model handled lighting, composition, and style.
That convenience made DALL-E one of the most accessible image generators ever built. But OpenAI has since folded image generation directly into its multimodal models. On March 25, 2025, GPT Image replaced DALL-E 3 as the default image generator in ChatGPT. The DALL-E 2 and DALL-E 3 models were removed from the OpenAI API on May 12, 2026, and the standalone DALL-E GPT was retired from ChatGPT on August 30, 2026. In other words, DALL-E as a standalone product is gone — its successor is ChatGPT Images, built on the GPT Image model family.
The current version, ChatGPT Images 2.6, launched on November 4, 2026, with roughly 50% faster generation, animated GIF creation (still rolling out), and saved reference images you can call up with @name. If you search for “DALL-E” in 2026, this is the tool you actually end up using.
Key Features
- Conversational image creation: describe an image in plain English, then refine it by chatting — “make it brighter”, “add a cat in the corner” — and the model revises the picture.
- Accurate text rendering: GPT Image renders readable text inside images far more reliably than DALL-E 3 did, making it practical for posters, labels, and thumbnails.
- Photo editing and inpainting: upload a photo and ask for changes in plain language, or highlight one area and describe the edit. Reference images keep a character, product, or style consistent across generations.
- Multiple aspect ratios: square, portrait, and landscape outputs for social posts, blog headers, and print.
- Images with thinking: on Plus and Pro plans, the model can reason about a complex brief before generating, producing more precise results for detailed instructions.
- Provenance metadata: generated images carry C2PA metadata identifying them as AI-generated, which matters for publishers and brands.
DALL-E Pricing (via ChatGPT)
There is no standalone DALL-E subscription. Image generation is bundled into ChatGPT plans, so the price you pay depends on the plan you already have. Prices below were verified on November 10, 2026 — always confirm the current figures on OpenAI’s pricing page, because plan names and limits change.
| Plan | Price | Image generation |
|---|---|---|
| Free | $0 | Limited, slower image creation |
| Go | $8/month | More generations than Free |
| Plus | $20/month | Higher limits, images with thinking |
| Pro | $200/month | Faster, unlimited image creation |
Developers can also use the API, where the gpt-image-1 model is priced per image from around $0.009, with a lighter gpt-image-1-mini tier from around $0.005 per image. That makes API access cheap for apps that generate images on demand, with no monthly subscription required.
Pros and Cons
Pros:
- Easiest image generator to start with — no prompt engineering needed
- Excellent at following detailed, plain-English instructions
- Strong text rendering for posters, thumbnails, and labels
- Conversational editing: refine images by chatting, not by re-prompting from scratch
- Included with a ChatGPT subscription many people already pay for
- Commercial use of outputs is allowed under OpenAI’s terms
Cons:
- DALL-E 3 itself is retired — you are really reviewing GPT Image, not a standalone DALL-E app
- Slower than dedicated generators: a generation can take one to two minutes
- Less artistic range than Midjourney for stylized or cinematic looks
- Free-tier limits are unpublished and throttled at busy times
- No fine-grained controls like seed locking or model training
Who Is DALL-E (ChatGPT Images) Best For?
ChatGPT Images is best for content creators, bloggers, marketers, and small business owners who want usable images without learning a new tool. If you already pay for ChatGPT Plus, it is the most convenient image generator you own: blog headers, social graphics, product mockups, and presentation visuals are all a sentence away. It is also the best choice when an image must contain exact text — a headline, a label, a sign — because its text rendering is among the most reliable available.
It is less suited to artists chasing a distinctive signature style, or to teams that need batch generation, API pipelines, or precise control over every pixel. For those jobs, see our roundup of the best AI image generators in 2026, and our head-to-head Midjourney vs DALL-E comparison.
FAQ
Is DALL-E still available in 2026?
No, not as a standalone product. DALL-E 3 was replaced in ChatGPT by GPT Image in March 2025 and removed from the API in May 2026. Image generation in ChatGPT today is ChatGPT Images, powered by the GPT Image models.
Is there a DALL-E 4?
No. OpenAI has never announced a DALL-E 4. The current engine is GPT Image 1 (and its 1.5 update), delivered inside ChatGPT.
Can I use DALL-E images commercially?
Yes. Under OpenAI’s terms, you own the images you generate and may use them commercially, including for client work. Generated images carry C2PA provenance metadata.
How much does DALL-E cost?
It is bundled into ChatGPT: free with limits, $8/month on Go, $20/month on Plus, and $200/month on Pro. Via the API, GPT Image starts at around $0.009 per image.
Is DALL-E better than Midjourney?
For following exact instructions and rendering text, yes. For artistic, cinematic, or highly stylized images, Midjourney still leads. The right choice depends on the job — see our Midjourney vs DALL-E comparison for the full breakdown.
Verdict
DALL-E the product is retired, but DALL-E the idea — image generation in plain English — lives on inside ChatGPT Images, and it is better than the original. GPT Image follows instructions more accurately, renders text cleanly, and edits photos conversationally. It is slower than specialist generators and less artistic than Midjourney, but for everyday content work it is the most convenient image tool most creators already have. If you pay for ChatGPT Plus, you already own one of the best general-purpose image generators of 2026.
Rating: 4.2/5. Excellent ease of use and text accuracy; docked a point for slower generation and a narrower artistic range than the specialists.
Prices and model status verified November 10, 2026 from OpenAI’s announcements and ChatGPT pricing pages. Confirm current pricing on the official site before you buy.