P-Image is a sub-1 second text-to-image model released by Pruna on December the 2nd.
Pruna is into shrinking and speeding up AI models. On their site they say they’re an “AI Optimization Engine” that builds systems to speed up other image or video tools.
P-Image stands out for speed, low price, image quality, how well it follows prompts, and how well it handles text in pictures. So it’s built for folks who need quick and good results.
It’s hosted by Pruna through an API.
P-Image isn’t built from zero. It’s based on past work from Black Forest Labs and Alibaba. Pruna took those earlier models, tweaked them, and rolled them into something new that runs faster and works better for pro use.
When tested, outputs most closely resemble Z-Image Turbo and the model came out a week after Z-Image was released.
If you'd like to access this model, you can explore the following possibilities:
