DALL·E
In plain wordsOpenAI's pioneering series of image generators, named after the painter Salvador Dalí and the robot WALL·E, which showed in 2021–2023 that AI could draw almost anything described in words.
A series of OpenAI text-to-image modelstext-to-image model, n.: An AI system that creates a picture from a written description, such as 'a fox reading a newspaper, in watercolor', usually within seconds.. The original DALL·E (January 2021) was a 12-billion-parameter transformertransformer, n.: The design behind today's chatbots and much of modern AI. It handles text by letting every word look at every other word and decide which of them matter for understanding it. that generated images autoregressivelyautoregressive image generation, n.: Making a picture one piece at a time, each piece predicted from those already made — the way chatbots write text word by word. as a single stream of 256 text tokens followed by 1,024 image tokens, produced by a discrete VAE that compressed 256 × 256 images to a 32 × 32 grid; it was trained on 250 million web image–text pairs, and its samples were reranked by a contrastive image–text model. DALL·E 2 (April 2022) switched to diffusion: a prior generated a CLIPCLIP, prop. n.: An OpenAI model that learned to match pictures with their captions from hundreds of millions of web examples, so it can judge how well any piece of text describes any image. image embedding from the caption and a diffusion decoder rendered it (the 'unCLIP' method), adding inpaintinginpainting, n.: Filling in a missing or erased part of a picture so it blends seamlessly with the rest — removing a stranger from a photo, or repairing a damaged one., outpainting and variations. DALL·E 3 (September 2023) improved prompt adherence by training on detailed synthetic captions and was built into ChatGPT. GPT-4o's native image generation replaced it in ChatGPT in March 2025; OpenAI removed DALL·E 2 and 3 from its API on 12 May 2026 and retired ChatGPT's DALL·E GPT on 30 August 2026.
Origin
Announced by OpenAI on 5 January 2021, with the paper 'Zero-Shot Text-to-Image Generation' by Aditya Ramesh and colleagues. The name blends the surrealist painter Salvador Dalí with Pixar's robot WALL·E. DALL·E 2's striking images made AI image generation a mainstream topic in 2022.
Usage note
The viral 2022 'DALL·E mini', later renamed Craiyon, was an independent open-source project unrelated to OpenAI. As of October 2026, OpenAI offered image generation through its GPT Image models rather than DALL·E.
Related
- text-to-image modeltext-to-image model, n.: An AI system that creates a picture from a written description, such as 'a fox reading a newspaper, in watercolor', usually within seconds.
- CLIPCLIP, prop. n.: An OpenAI model that learned to match pictures with their captions from hundreds of millions of web examples, so it can judge how well any piece of text describes any image.
- autoregressive image generationautoregressive image generation, n.: Making a picture one piece at a time, each piece predicted from those already made — the way chatbots write text word by word.
- VQ-VAEVQ-VAE, n.: A network that compresses an image or sound into a short list of entries from a learned 'codebook' — like describing a picture with a fixed set of tiles — which other models can then generate.
- diffusion modeldiffusion model, n.: An AI that learns to create images, sound or video by practicing noise removal: shown pictures buried under static, it learns to clean them up, then builds new ones starting from pure static.
- MidjourneyMidjourney, prop. n.: A popular paid AI image generator, first used by typing prompts into the chat app Discord, known for its polished, painterly look; it can also animate images into short videos.
- Stable DiffusionStable Diffusion, prop. n.: A widely used image-generating AI whose model files were released for anyone to download, so people could run it on their own computers and adapt it.
Sources
- Ramesh, A. et al. (2021). Zero-Shot Text-to-Image Generation. ICML 2021. arXiv:2102.12092. arxiv.org/pdf/2102.12092v1.pdf
- Ramesh, A., Dhariwal, P., Nichol, A., Chu, C. & Chen, M. (2022). Hierarchical Text-Conditional Image Generation with CLIP Latents. arXiv:2204.06125. arxiv.org/abs/2204.06125
- OpenAI (2026). Deprecations. OpenAI developer documentation. developers.openai.com/docs/deprecations
Mentioned in
AI artAI art, n.: Images, music or other artworks made with the help of AI — from early computer drawing programs to today's prompt-based image generators — and the debate over whether and how such work counts as art., MidjourneyMidjourney, prop. n.: A popular paid AI image generator, first used by typing prompts into the chat app Discord, known for its polished, painterly look; it can also animate images into short videos., OpenAIOpenAI, prop. n.: The American company that makes ChatGPT and the GPT models. It began in 2015 as a nonprofit research lab and is now a for-profit company overseen by a nonprofit foundation., text-to-image modeltext-to-image model, n.: An AI system that creates a picture from a written description, such as 'a fox reading a newspaper, in watercolor', usually within seconds., ImagenImagen, prop. n.: Google's family of AI image generators, first shown in 2022, whose research showed that better understanding of language yields pictures that match descriptions more faithfully., Mira MuratiMira Murati, prop. n.: An Albanian-born American engineer who, as OpenAI's chief technology officer, oversaw the launch of ChatGPT, and who founded the AI company Thinking Machines Lab in 2025., SoraSora, prop. n.: OpenAI's video-generating AI, unveiled in 2024 with startlingly realistic clips; a 2025 version added sound and a social app, but OpenAI shut the service down in 2026., autoregressive image generationautoregressive image generation, n.: Making a picture one piece at a time, each piece predicted from those already made — the way chatbots write text word by word., Gumbel-softmaxGumbel-softmax, n.: A mathematical trick that lets a neural network make a choice among separate options, such as picking one word or one category, while still learning through small, smooth adjustments., VQ-VAEVQ-VAE, n.: A network that compresses an image or sound into a short list of entries from a learned 'codebook' — like describing a picture with a fixed set of tiles — which other models can then generate.