OpenAI Expands Image Generation Access to All Users
OpenAI has made its advanced image generation technology available to all users, including those on free plans, CEO Sam Altman announced via social media. Previously, the feature was exclusive to paying subscribers of the company's ChatGPT service.
While free tier users can now access the image generator, powered by OpenAI's GPT-4o model, it appears to be limited to creating three images daily.
The newly accessible tool offers several technical advancements over previous image generators. According to the company, the system was trained on the "joint distribution of online images and text," allowing it to understand not just how images relate to language but how images relate to each other.
Key capabilities of the new image generator include:
Accurate text rendering within images
Multi-turn generation that maintains consistency across iterations
Advanced instruction following with up to 20 different objects in a single prompt
Integration with ChatGPT's existing knowledge base
Ability to transform or draw inspiration from uploaded images
It can make fake documents like restaurant receipts
The technology's rapid adoption has already sparked several creative and controversial uses:
Studio Ghibli-style animations raising copyright questions
Fake documents like restaurant receipts
On the positive side:
Detailed infographics with precise text placement
Character design iterations maintaining consistency
Diagrams and technical illustrations
Will take actions:
Responding to potential misuse concerns, an OpenAI spokesperson told TechCrunch that all generated images contain metadata identifying them as ChatGPT creations. The spokesperson added that the company "takes actions" when images violate its guidelines.