OpenAI Unveils DALL-E 2: 4x the Resolution and a Major Leap in Realism

OpenAI announced “DALL-E 2,” an AI system that generates realistic images and art from text descriptions. With four times the resolution of the original DALL-E announced in January 2021, and 88.8% of human evaluators favoring it for realism in side-by-side comparisons, it drew widespread attention as a model that dramatically raised the practical bar for image generation AI.

Details

• Performance improvement: Generates images at four times the resolution of the original DALL-E, with greater accuracy and realism
• Evaluation results: In comparisons by human evaluators, 71.7% favored DALL-E 2 for caption matching and 88.8% favored it for realism
• Phased rollout: At launch, access was limited to a small group of trusted invited users, with a cautious rollout approach
• Safety measures: The design emphasized excluding explicit sexual content from training data, safeguards around facial recognition, and mechanisms for monitoring misuse

What happened next

DALL-E 2 expanded to an invite-based beta in July of the same year, and by September the waitlist was removed, allowing anyone to use it. Alongside Midjourney and Stable Diffusion, this announcement was one of the events that brought image generation AI into mainstream awareness in 2022.