SAN FRANCISCO — OpenAI‘s DALLĀ·E beta is available to everyone without a waitlist. New users can sign up and begin creating immediately. This decision comes after extensive testing and enhancements to the safety systems, allowing for broader access. Currently, over 1.5 million users are generating more than 2 million images daily with DALLĀ·E, ranging from artists to architects.
Illustration: Justin Jay Wang Ć DALLĀ·E
Expanding Access and Enhancing Creativity
Since its initial preview in April, DALLĀ·E has proven to be a versatile tool for various creative fields. The feedback from users has been instrumental in developing new features, such as Outpainting, which allows users to extend images beyond their original borders, and collections to streamline creative processes.
Learning and Improving
The iterative deployment approach has enabled OpenAI to refine DALLĀ·E’s safety systems. The enhancements include more robust filters to reject inappropriate content and advanced detection techniques to prevent misuse. These improvements make the system safer for a wider audience.
Upcoming Developments
OpenAI is also testing a DALLĀ·E API with select customers and plans to expand access to developers and businesses soon. This will enable the integration of DALLĀ·E’s capabilities into various applications, broadening its utility.
Join the Creative Community
The journey of DALLĀ·E over the past few months highlights its growth and the expanding community of users. Notable milestones include the addition of the Outpainting feature, the launch of the beta with pricing, and the publication of significant research papers.
Sign up today to start exploring the creative possibilities with DALLĀ·E.
Key Milestones
- September 2022: DALLĀ·E beta is available without a waitlist, with over 1.5 million users creating more than 2 million images per day.
- July 2022: Outpainting feature added.
- June 2022: Beta version launched with pricing, onboarding over 1 million users.
- April 2022: Research preview expanded to 1,000 users per week.
- March 2022: DALLĀ·E 2 paper published, research preview opened to 200 users.
- November 2021: First version of GLIDE paper published.
- April 2021: Diffusion models outperformed GANs on ImageNet synthesis.
- December 2020: OpenAI announced the first DALLĀ·E and open-sourced CLIP.