OpenAI has unveiled an upgraded iteration of its image generation feature on ChatGPT, now powered by the enhanced GPT Image 1.5 model. This new model boasts the ability to make more precise alterations to uploaded images while retaining essential aspects such as lighting and composition. Additionally, it can produce images up to four times faster than its predecessor.
OpenAI claims that the updated model excels in various tasks, including executing intricate instructions, rendering text, and managing edits that involve adding or removing elements from an image. To complement this model enhancement, OpenAI has introduced a dedicated section for image creation within the ChatGPT interface. This section offers preset filters and prompts to assist users in initiating image generation without the need for detailed descriptions.
The updated model has been gradually rolled out to all ChatGPT users since December 16, with availability for Business and Enterprise clientele anticipated at a later date. Furthermore, developers can access the model through an API.
While acknowledging that the update signifies a significant advancement, OpenAI acknowledges that there is room for further enhancements as results are not flawless. The latest API version, GPT Image 1.5, comes with a 20% reduction in pricing compared to its predecessor for both image inputs and outputs.
OpenAI highlights that businesses in creative tools, e-commerce, and marketing software sectors are already leveraging the new model for various applications, such as creating product image collections and maintaining brand visuals.