OpenAI has released ChatGPT Images 2.0, an update to its image generation model that incorporates reasoning capabilities. The new version allows the model to think through prompts iteratively, improving accuracy in tasks such as rendering text within images and following complex instructions.
According to OpenAI, Images 2.0 is built on the GPT-4o architecture and can generate images with better text rendering, more precise object placement, and improved adherence to user prompts. The model can also refine images through a multi-step reasoning process, similar to how ChatGPT's text models handle complex queries.
The update is available to ChatGPT Plus, Pro, and Team users, with enterprise access coming soon. OpenAI has not specified a release date for free-tier users. The company emphasized that the model is designed to reduce common errors like misspelled words or distorted objects.
This release follows OpenAI's earlier image generation capabilities introduced in 2024. The company has not disclosed specific performance metrics or comparisons with previous versions.
