Originally posted on Linkedin on Apr 25, 2025
The new image generation tool of OpenAI we’ve all seen the past weeks, is now available on their API as gpt-image-1. It’s slow. It’s expensive (at ~25ct per image). But now that it works under their regular API rate limiting scheme, it can reliably churn out images at a continuous rate.
While there may be alternatives available on the market who are faster and cheaper, I have not found one with such amazing prompt adherence as gpt-image-1.
Prior, applications were limited to diffusion-style image generation techniques, which are great for creative art; now it’s useable for informative images. Think of business infographics, generated restaurant menu’s, public signage, educational images. For especially education I expect a huge push in the near future: on-demand, situational, level-adjusted explanative imagery sounds amazing to have as a teacher.
Or perhaps just some more plastic action figures.