OpenAI has released a new version of its image generation feature within ChatGPT, which is powered by an updated model called GPT Image 1.5.

The company states the new model is capable of making more precise edits to uploaded images while preserving key details like lighting and composition. It is also designed to generate images up to four times faster than its predecessor. 

According to OpenAI, the model performs better at tasks including following complex instructions, rendering text, and handling edits that add or remove elements from a picture.

Alongside the model update, OpenAI is introducing a new dedicated space for image creation within the ChatGPT interface. This area will provide preset filters and prompts to help users start generating images without having to write a detailed description. 

The model began rolling out to all ChatGPT users on December 16, with availability for Business and Enterprise customers expected later. It is also accessible to developers through an API.

OpenAI acknowledged that while the update represents a "meaningful step forward," results remain imperfect and there is significant room for future improvement. The new API version, GPT Image 1.5, is priced 20% lower than the previous model for image inputs and outputs.

The company noted that businesses in creative tools, e-commerce, and marketing software are already using the new model for tasks like creating product image catalogues and maintaining brand visuals.