What is DALL-E
DALL-E is OpenAI’s earlier text-to-image model family. OpenAI’s current model catalogue labels DALL-E 3 and DALL-E 2 as deprecated and recommends the newer GPT Image family for current image-generation and editing work, so buyers should treat DALL-E as a legacy API option rather than OpenAI’s forward-looking image platform.
Core offerings
- DALL-E 3 image generation from natural-language prompts through the Images API.
- Standard and HD quality options at square, portrait, and landscape sizes.
- Text input and image output; DALL-E 3 does not support image input or image editing.
- GPT Image 2 as OpenAI’s current state-of-the-art generation and editing model.
Pricing
As of July 20, 2026, OpenAI’s official DALL-E 3 model page lists Standard quality at US$0.04 per 1024×1024 image and US$0.08 per 1024×1536 or 1536×1024 image. HD quality is US$0.08 per 1024×1024 image and US$0.12 per portrait or landscape image.
DALL-E is usage-priced rather than sold in named subscription tiers. OpenAI’s model catalogue marks both DALL-E models as deprecated and identifies GPT Image 2 as the current image model, so new integrations should verify the current GPT Image pricing and migration path before choosing a legacy DALL-E endpoint.
Model footprint
DALL-E 3 remains documented as a previous-generation image model, but it is no longer OpenAI’s recommended model family. It generates new images from text; teams needing image inputs, iterative edits, or OpenAI’s latest quality should assess GPT Image 2.
Why select DALL-E
DALL-E 3 is relevant mainly for maintaining an existing integration with predictable per-image rates. New production work should normally start with OpenAI’s current GPT Image documentation.
Official sources: DALL-E 3 model and pricing, OpenAI model catalogue.