If you have ever used an AI text-to-image generator, you are likely familiar with DALL-E, OpenAI’s model that originally sparked the image generator interest. Last month, OpenAI announced an even more advanced DALL-E, and now it’s here.
On Monday, OpenAI shared via its release notes that DALLE-3 is rolling out in beta, making DALL-E 3 available directly from ChatGPT on web and mobile for select users.
Also: China’s Ernie AI is just as good as – or better than – ChatGPT, say its creators
Although the release notes didn’t specify who DALLE-3 was rolling out to, when OpenAI originally announced the model in September, the company shared it would be rolling out to ChatGPT Plus and Enterprise customers in October.
To access DALL-E 3 in ChatGPT, users will have to choose DALLE-3 from the dropdown selector under the GPT-4 option.
Once selected, users can type in text descriptions as short as a sentence or as detailed as a paragraph to describe the new art they want to see depicted right into ChatGPT, where DALLE-3 will generate the image in seconds.
When first announced, OpenAI shared that DALLE-3 would be able to surpass DALLE-2 capabilities by creating more accurate images and better capturing the prompt’s nuances.
Also: DALL-E 3 is now available for free in Bing Chat
If you want to test the difference for yourself, DALLE-3 is already available in Bing Chat for free. As seen by ZDNET’s experience with DALLE-3 in Bing Chat, the images rendered are of higher quality, with more attention to detail.
The release notes also revealed that Browsing, which relaunched at the end of September after being removed from the platform due to user misuse, is moving out of beta.
This means that Plus and Enterprise users will be able to choose “Browse with Bing” from the GPT-4 model selector without having to switch the beta toggle.