DALL-E

DALL-E was first introduced by OpenAI in January 2021 as a text-to-image generative model. The name is a portmanteau of the surrealist artist Salvador Dalí and the Pixar character WALL·E, reflecting the model's ability to generate creative and sometimes surreal images from text descriptions.

DALL-E was introduced in January 2021, DALL-E 2 followed in April 2022 with significantly improved image resolution and the addition of inpainting and outpainting capabilities, and DALL-E 3 was released in 2023 with integration directly into ChatGPT. Each successive version improved text rendering accuracy and the ability to follow complex, multi-element prompts.

DALL-E uses a transformer-based architecture and is built on technology related to OpenAI's GPT models, trained to generate images rather than text from input prompts. DALL-E 2 incorporated CLIP (Contrastive Language-Image Pre-training) for improved alignment between text descriptions and generated images.

DALL-E 2 was made available to the public through OpenAI's web interface after an initial waitlist period. DALL-E 3 is accessible to users through ChatGPT Plus subscriptions and through Microsoft's image creation tools integrated into Bing and other Microsoft products.