OpenAI has been in a giving mood lately, offering free plan users more and more features that were previously reserved for the Pro version of ChatGPT. Now it’s the turn of AI image generation in the form of access to DALL-E 3.
DALL-E 3 is OpenAI’s AI image generator and, similar to tools like Midjourney, Flux or Ideogram, can create photorealistic images, render text and enable inpainting.
With the launch of GPT-4o earlier this year, OpenAI made most of its paid features, including image analysis, encoding, and the use of custom GPTs, free for the first time. GPT-4o, the platform’s most powerful model, was even made available to free users (with a daily usage limit), but image generation was not on that list.
How to create free images with ChatGPT?
It’s straightforward to use, though, as ChatGPT does most of the work. As OpenAI wrote on X, “Simply ask ChatGPT to create an image for a slide deck, personalize a card for a friend, or show you what something looks like,” and it will do it for you.
Be as verbose as possible, otherwise ChatGPT will fill in all the gaps based on its own training data. So if you want the card to have a cat on it, tell it that. The text representation isn’t perfect yet either, but if you put the words in quotes it will be more accurate.
While free version users can now create images, it’s limited to two per day, so if you want to create multiple revisions or even enough images for a comic book, you’ll need to find another platform or pay $20 per month for the Pro version.
What else is free in ChatGPT?
We’re introducing the ability for ChatGPT Free users to create up to two images per day using DALL·E 3. Just ask ChatGPT to create an image for a slide deck, personalize a card for a friend, or show you what something looks like. pic.twitter.com/3csFTscA5I8 August 2024
Earlier this year, OpenAI announced its goal of making the most powerful artificial intelligence technology available to as many people as possible.
To achieve this goal, the company provided free users with many of the features that were standard for Pro subscribers, including the ability to use a number of custom chatbots and third-party-created features.
The biggest change was that we finally had access to the most powerful model – in this case GPT-4o – rather than using the two-year-old GPT-3.5. That model has since been beaten in benchmarks by models that can run on a smartphone.
Free users also have access to the upload feature. This allows them to use GPT-4o’s AI vision and analytics capabilities, which allow it to look at the details of an uploaded image and outline what it contains. I like to use this the most to enter ingredients and have it suggest a recipe.