Content
summary Summary

OpenAI is testing a new user interface for DALL-E 3.

You can choose between predefined styles and an aspect ratio directly in the DALL-E 3 GPT: Widescreen, 1:1 or Vertical.

The option appears above the chat input line. Clicking on an item will insert it into the chat input line as a text item. Hovering over a style will visualize it with a cat image in that style. This is a useful educational aspect, as most users probably don't know all the art styles out there.

Image: Screenshot of THE DECODER

Up to now, OpenAI has completely renounced control options outside of text input for DALL-E 3 and offers significantly less features than the competition from Midjourney, Stable Diffusion or Ideogram.

Ad
Ad

The new feature does not change this fact, but is at least an indication that OpenAI is still working on the DALL-E interface.

DALL-E 2 had a graphical user interface that also offered functions such as outpainting and inpainting, i.e. adding elements to an image or changing an element in the image using text prompts.

DALL-E 3 offers more visual variety via the API: There are two basic modes, "Natural" and "Vivid," which generate either natural, realistic or hyper-realistic, dramatic content.

Two other quality modes are "HD" and "Standard". HD mode is designed to display more detail and follow prompts more accurately, but is more expensive and takes longer.

DALL-E 3 follows prompts more accurately than comparable systems. However, Midjourney and Ideogram in particular have caught up with the latest updates and often offer better image quality, especially Midjourney, and significantly more setting options. Ideogram can render fonts better than Midjourney and DALL-E 3. The new Stable Diffusion 3 might follow text better.

Recommendation

One benefit of DALL-E: OpenAI has implemented the C2PA standard, which embeds metadata in the generated images to verify their origin. The metadata can be read by certain websites and social media platforms. OpenAI is also working on an image classifier that reliably recognizes DALL-E images.

Ad
Join our community
Join the DECODER community on Discord, Reddit or Twitter - we can't wait to meet you.
Ad
Join our community
Join the DECODER community on Discord, Reddit or Twitter - we can't wait to meet you.
Support our independent, free-access reporting. Any contribution helps and secures our future. Support now:
Bank transfer
Summary
  • OpenAI is testing a new user interface for DALL-E 3 that allows you to choose between predefined styles and aspect ratios (widescreen, 1:1 or vertical).
  • The options appear above the chat input line. Until now, DALL-E 3 had fewer features than its competitors Midjourney, Stable Diffusion or Ideogram. The new feature doesn't change that, but it shows that OpenAI continues to work on the user interface.
  • Compared to other image AI systems, DALL-E 3 is more accurate at following prompts, but Midjourney and Ideogram offer more image quality and setting options.
Online journalist Matthias is the co-founder and publisher of THE DECODER. He believes that artificial intelligence will fundamentally change the relationship between humans and computers.
Join our community
Join the DECODER community on Discord, Reddit or Twitter - we can't wait to meet you.