OpenAI CEO Sam Altman has announced the DALLE-2 successor via Twitter. We predict AI-generated images are going to be indistinguishable from reality in 2025, largely thanks to DALL-E 3 (Stylized DALL·E 3). In the meantime, what is our DALL-E 3 release prediction?

DALL-E 3 release window prediction

Announced on August 20th, 2023, DALL-E 3 is officially here! DALL-E 3 (Stylized DALL·E 3) has now rolled out to all paid users across the ChatGPT Plus, Teams, and Enterprise plans. This means that ChatGPT users with a paid subscription can generate unlimited images with OpenAI's AI art generator at no additional cost. The AI image generator expands the capabilities of ChatGPT into the visual realm, making it a multimodal generative AI model for the first time.

OpenAI is also working on an audio AI called Whisper, which powers the speech-to-text functionality on mobile, and may soon power a text-to-speech function. We predict this will all be possible by Q3 2024.

We predicted that DALL-E 3 would be released (meaning available to the public) in late Fall, 2023. The completion of the rollout was announced via Twitter on August 19th, 2023.

DALL·E 3 is now available to all ChatGPT Plus & Enterprise users, letting you create unique images through conversation. Describe your vision, let ChatGPT generate multiple variants, and then request edits — all in real-time.

OpenAI, X (formerly Twitter)

Available to the public does not mean available for free, however. While ChatGPT will maintain a free version forever, DALL-E has never been free, and instead relies on a paid token system. Essentially, DALL-E uses a pay-as-you-go (PAYG) business model.

Essential AI Tools

Is DALL·E 3 available now?

Yes! DALL·E 3 is available everywhere, as the rollout is now complete. You can use the AI image generator today, through ChatGPT and Bing Image Creator.

Will DALL·E 3 be available to the public?

Yes, DALL·E 3 is available to the public through a paid subscription to either ChatGPT Plus and ChatGPT Enterprise. Since the release of DALL-E 2 in September 2022, leading competitors Midjourney and Stability AI (creator of Stable Diffusion) haven’t had much to worry about. Now though, recent breakthroughs in image fidelity have lead to fair comparisons between DALL·E and Midjourney, with text-in-image generation far surpassing either of the rivals.

Aditya Ramesh, head of the DALL-E team, assures us that prompt engineering will also be a thing of the past. “You won’t really have to worry about fussing around with really long prompts,” explains the OpenAI lead researcher. “Instead, you can just interact with ChatGPT as if you were talking to a coworker.”

Safety features preventing misinformation and hateful content

OpenAI claims it’s new AI art generator will feature robust safety measures to prevent the generation of lewd or potentially hateful images. DALL·E 3 is reportedly unable to recreate images of public figures where the text prompt explicitly mentions them by name.

Sandhini Agarwal, AI Policy researcher at OpenAI, reports “high confidence” in its safety measures, and continual focus on improvement. While DALL·E 2 has always been copyright-free, that doesn’t mean it is free from guilt. The predecessor model is able to mimic the style of human artists, but sources say that DALL·E 3 is restricted from doing so, which safeguards creators illustrations, and ultimately their livelihood!

When can you access OpenAI’s latest AI model?

DALL·E 3 is available now.

DALL·E 3 is now in research preview, and will be available to ChatGPT Plus and Enterprise customers in October, via the API and in Labs later this fall.

OpenAI.com

It’s predecessor, DALL·E 2, is still available for anyone to use, via desktop or mobile, and can do much the same thing as DALL·E 3 (with less photo realism and text-in-image accuracy). The prior model already has a ChatGPT plugin of its own, meaning that you can use both within ChatGPT.

FAQ

Do I need ChatGPT to get DALL·E 3?

No, you don’t need ChatGPT to use DALL·E 3. You can use the AI image generator via the ChatGPT interface (via web browser or ChatGPT app), or Bing Image Creator, or the OpenAI Labs playground (via web browser), or the OpenAI API. Regardless of where you use it, you will need to pay for it.

ChatGPT itself is not considered a competitor to DALL·E 3. They’re both currently supported models from the same company (OpenAI). ChatGPT is an AI chatbot, which uses a variety of neural networks, with the current most powerful option being GPT-4. This is a large language model (LLM) which will handle the natural language processing (NLP) of your text prompts for DALL·E 3, regardless of where you choose to access it. Large language models purely process text data. GANs (Generative Adversarial Networks) like the one which powered DALL·E 1, and diffusion models like the ones which power DALL·E 2, and DALL·E 3, process images. The two can be combined, as seen in the integrated of DALL·E 3 within the ChatGPT interface!

How much do DALL-E 3 credits cost?

Each DALL·E user will receive “50 free credits during their first month of use and 15 free credits every subsequent month”. AI image generations cost one credit each  —  which provides 4 options to choose from  —  “or an edit or variation prompt, which returns three images.”

Free credits are available to early adopters who signed up to use DALL·E before April 6, 2023.

OpenAI.com

Use cases for OpenAI’s AI image generator

The firm makes its own suggestions via the DALL-E 3 rollout announcement. Here are some ideas for prompts, and their results!

Prompt: “I am doing a report on cirrus clouds for my science class. I need photorealistic images that show off how wispy they are. I am going to compare them to photos I took of puffy cumulonimbus clouds at my house yesterday.”

Prompt: “I am working with a web designer for a new website to sell bicycle helmets. Could you recommend a few website UI’s that are both modern and simple that I can share with her as examples?”

Prompt: “My business partner and I are creating a coffee shop inspired by our mutual love for rabbits. We have a logo we like a lot that is minimalist with a rabbit, but anxious to get your take on it.”