Currently, generating images in ChatGPT using Dall-E 3 may result in outputs that appear overly cartoonish with exaggerated details, as seen in the example below.
While these outputs are decent, they fall short of perfection due to their distinctly AI-generated look. The good news is that OpenAI plans to introduce 4o Image Generation, which promises to deliver more natural, dynamic, and significantly improved images compared to the current technology.
For instance, the image above demonstrates a more realistic and refined result, a noticeable improvement over the current technology’s output.
Additionally, users will soon be able to create images with embedded text descriptions. For example, a command like the following will produce a highly detailed and visually appealing image:
As shown in the image above, users will also have the ability to add text details to their creations, making the experience even more impressive.
OpenAI has also showcased how their AI can be used for educational purposes. For example, users can generate images depicting marine life, as illustrated in the following example.
This is truly remarkable, as GPT-4o will not only enhance search results but also elevate image quality to a level where it may be indistinguishable from human-taken photos.
Look at the image below—it’s incredibly realistic and downright stunning.
However, it’s important to note that achieving such high-quality results comes at a cost. Access to GPT-4o requires a subscription to ChatGPT Plus or ChatGPT Pro, priced at $200 USD per month, offering extensive and powerful features.
What are your thoughts? Share your comments below.
Via : OpenAI
