OpenAI’s New Image Model is a Response to the AI Creative Race

OpenAI recently called a Code Red!, which most people thought was about how Gemini was becoming really good at AI tasks. This new version of ChatGPT Images, which is powered by the GPT Image 1.5 model, could be the product that was released after this CODE RED!. Because if you have seen Gemini’s Banana Pro, it is really great at making real-looking images.

One cool thing that this new Image model can do is change only what is asked for, while keeping lighting, composition, and people’s faces consistent. It can create 6x6 grids and render dense, small text even more accurately than the older version. To jumpstart inspiration, it has a dedicated Creation space that allows you to pick from preset filters and prompts, Likeness upload to reuse your appearance across different images without having to re-upload photos, and Creative transformation that lets you turn photos into specific styles, like a 90s movie poster or comic style, while keeping the original way it looked.  Oh, it is also 4x faster and will be available to all users, even free ones.

Image’s model is 20% cheaper for developers, making it great for e-commerce brands. This model is obviously still not perfect. OpenAI even acknowledges some inaccuracies and challenges with multilingual text in a blog post. This update is only the beginning of a shift towards more natural-looking AI creative tools from OpenAI. So we’ll have to see if any other features come out of the CODE RED!.

Do you like all the new AI features that have been coming out lately? Do you generate a lot of images with AI? Do you think that ChatGPT is losing relevance or becoming even more popular? Let us know in the comments!

Source and Images: ChatGPT