What is DALL-E?

Updated 2026-07-152,400 searches/mo· DALL-E
Short answer

DALL-E is OpenAI's text-to-image AI that turns written prompts into original pictures, first announced in January 2021. As of July 2026 the standalone DALL-E 2 and DALL-E 3 models are retired; OpenAI replaced them with its GPT Image models, which now power image generation inside ChatGPT.

Why — the first-principles explanation

DALL-E is a text-to-image model. You type a description like "a corgi wearing sunglasses on a skateboard," and it paints a brand-new picture that matches. It learned to do this by studying hundreds of millions of image-plus-caption pairs from the web, building a statistical map between words and the visual patterns those words describe.

Under the hood, the newer versions use diffusion. The model starts with a field of random noise, like TV static, then removes that noise step by step, nudging every pixel toward something that fits your words. A language component reads your prompt and steers each denoising step, which is why longer, more specific prompts produce closer matches.

The name is a pun on the surrealist painter Salvador Dalí and Pixar's robot WALL-E. DALL-E (2021), DALL-E 2 (2022), and DALL-E 3 (2023) each got sharper and better at following instructions. In 2025-2026 OpenAI folded image generation directly into its main GPT models, so the DALL-E brand was retired and replaced by GPT Image.

An example that makes it click

Imagine a lightning-fast sketch artist who has read every picture book and photo caption ever printed. You whisper "a red bicycle floating in space next to a slice of pizza," and instead of copying an existing drawing, they invent a fresh one on the spot that fits your exact words. DALL-E is that artist, except it draws by cleaning up a page of random static until the picture appears.

Key facts

Infographic: What is DALL-E — short answer and key facts
Visual summary — What is DALL-E?
DE
Try DALL-E by OpenAI

OpenAI's image generator, built into ChatGPT.

Affiliate link — we may earn a commission at no cost to you.
Visit DALL-E ↗
▶ The 60-second explainer (script)

DALL-E is OpenAI's text-to-image AI: you type a sentence, and it generates a brand-new picture to match. It was announced in January 2021 and got its famous name from painter Salvador Dali and the robot WALL-E. Here's the trick: it doesn't search for an existing image. It starts with random static, then removes the noise step by step, guided by your words, until a clear picture forms. That's called diffusion. Over time DALL-E 2 and DALL-E 3 got sharper and better at following instructions, and DALL-E 3 was built right into ChatGPT. One important update: as of 2026, OpenAI has retired the standalone DALL-E models and replaced them with newer GPT Image models. So when you make pictures in ChatGPT today, you're using GPT Image, the successor that grew out of DALL-E.

What authoritative sources say

OpenAI API Deprecationsofficial — OpenAI retired DALL-E 2 and DALL-E 3 from its API on May 12, 2026, recommending GPT Image models as replacements. source ↗
Wikipedia: DALL-Eorg — DALL-E is a text-to-image model first announced in January 2021, named after Salvador Dali and WALL-E. source ↗

People also ask

Is DALL-E still available in 2026?

Not as a standalone model. DALL-E 2 and DALL-E 3 were retired in May 2026 and replaced by OpenAI's GPT Image models, which now handle image generation in ChatGPT.

What does DALL-E stand for?

It isn't a formal acronym. The name is a playful blend of surrealist painter Salvador Dali and Pixar's animated robot WALL-E.

Who made DALL-E?

OpenAI, the same company behind ChatGPT. DALL-E was one of its earliest public image-generation projects, announced in January 2021.

Is DALL-E the same as ChatGPT?

No. ChatGPT is a chat assistant, while DALL-E was the image model it called. Today ChatGPT generates images using GPT Image, DALL-E's successor.

Related questions