OpenAI has released a new version of ChatGPT Images, introducing an updated image generation and editing experience within ChatGPT. The update is powered by a new flagship image model and is designed to improve how users create and modify images, whether starting from scratch or editing existing photos.
The new model focuses on more accurate instruction following, allowing users to make specific changes while keeping important details such as lighting, composition and facial features consistent. OpenAI says image generation is now up to four times faster, making it easier to iterate on ideas without long wait times. The model supports a wide range of edits, including adding, removing, blending and transforming elements within an image.
A new Images space has also been added to ChatGPT, offering preset styles and prompts to help users explore creative ideas without writing detailed instructions. Improvements have also been made in text rendering, enabling clearer handling of smaller and denser text within images.
The updated image model is rolling out to all ChatGPT users and is also available through the API as GPT Image 1.5. Business and Enterprise access to the new Images experience will follow later.
Leave a comment