MUMBAI: OpenAI has introduced ChatGPT Images 2.5, a new image generation model designed to deliver sharper details, faster generation and more precise editing across creative workflows.
OpenAI said people create more than 3 billion images every week across ChatGPT Images and its GPT-Image models in the API. Images 2.5 builds on the previous generation with improvements in image quality, reference handling and multi-step editing.
The model produces more natural lighting and richer textures while improving its ability to preserve people and other subjects from reference photographs. It is also designed to follow editing instructions more reliably across multiple turns.
Generation latency has been reduced by up to 50 per cent compared with Images 2.0, allowing users to create and refine images faster.
OpenAI is also introducing new features in ChatGPT to give users greater control over the creative process.
Sketch allows users to draw directly in ChatGPT and use the drawing as a reference for the final image. Templates are being added for common formats such as flyers and product photographs, making it easier to start new projects.
Users can also place comments directly on images to guide more focused edits. OpenAI is additionally allowing users to share the prompts used to create images, enabling others to apply the same ideas to their own photos and details.
ChatGPT Images 2.5 is available to ChatGPT, ChatGPT Work and Codex users across desktop, mobile and the web.
For developers, OpenAI is introducing two models through its API.
GPT-Image-2.5 Flare brings the improvements in image quality, editing and generation speed, while GPT-Image-2.5 Sunburst is designed for more detailed creative work, offering greater precision at the cost of longer generation times.
OpenAI said Images 2.5 is better at using reference photographs to transform familiar subjects into new settings, visual styles and compositions.
The model is designed to make subjects more recognisable while preserving distinctive features, with improvements to lighting and textures. For developers using reference-led workflows, this is intended to make image variations more reliably anchored to the original source.
The company said the improvements are aimed at making image generation more useful not only for one-off creations but also for iterative creative workflows where users repeatedly refine an image.
