DALL-E 3
DALL-E 3 is a historical OpenAI text-to-image model. OpenAI’s deprecation record states that dall-e-3 was removed from the API on May 12, 2026. Keep this page as a reference for old prompts, tutorials, and integrations. It is not a current free service or a model to choose for a new integration.
What it offers
DALL-E 3 generated images from written descriptions, with parameters specific to that model. Old instructions for its quality, style, size, and URL responses should not be copied unchanged to GPT Image. DALL-E 3 generation also should not be confused with image edits or variations supported by other image models. A shared Images API does not mean identical model capabilities.
Access and cost
OpenAI’s retirement table lists GPT Image replacements. Its current image guide recommends the GPT Image 2.5 models for new integrations and documents generation and editing through supported APIs. For interactive use, start with the current ChatGPT image experience instead of assuming an old DALL-E integration still exists. A ChatGPT subscription does not pay for API usage, which has its own model and token charges.
Practical limits
When migrating, update the model identifier, supported parameters, output handling, and cost calculation together. Preserve prompts and saved images before replacing an old workflow. Compare representative outputs for composition, typography, safety behavior, and latency. The replacement model will not reproduce the original picture exactly, and old per-image prices are not a quote for current GPT Image models.