OpenAI supercharges ChatGPT images — Here’s what the new GPT Image 1.5 can do

On December 17, OpenAI introduced a significant upgrade to its ChatGPT Images, debuting the advanced GPT Image 1.5. This release transforms the way users create and modify images with artificial intelligence, providing a global rollout for all ChatGPT users and API clients. The latest model promises a remarkable enhancement in speed, generating images up to four times faster while ensuring that crucial aspects like lighting, composition, and the appearance of subjects remain consistent during edits. Users can now generate images from scratch or refine existing photos with improved accuracy. A standout feature of GPT Image 1.5 is its new creative workspace within ChatGPT, which simplifies image exploration through preset styles, filters, and prompts. Additionally, the model exhibits improved instruction-following capabilities. When modifying an uploaded image, it focuses solely on the requested changes, leaving the rest intact. This leads to more realistic edits, effective virtual clothing and hairstyle try-ons, as well as cleaner stylistic filters and conceptual transformations. With robust editing functionalities, users can add, remove, blend, combine, and transpose elements without compromising the quality of the original image. OpenAI highlights the model’s prowess in creative transformations, enabling the addition of text, layout adjustments, and reimagining visuals while preserving the essence of the original content. The text rendering enhancements further allow for more precise handling of smaller and denser text. Moreover, improvements in processing multiple small faces and achieving more natural outputs enhance the usability of images in real-world applications. The advancements in the API for GPT Image 1.5 also focus on better image preservation and brand consistency, making it highly beneficial for marketing, logo design, and e-commerce product catalogs generated from a single source image. In an effort to make the technology more accessible, OpenAI has reduced image input and output costs by 20% compared to its predecessor, facilitating greater image generation within existing budgets for developers and businesses. Enterprises and startups in creative tools, e-commerce, and marketing sectors are already leveraging GPT Image 1.5. Developers can explore the new model in the OpenAI Playground or refer to the official prompt guide for best practices. The upgraded ChatGPT Images is now available worldwide, with no manual model selection required. The previous version will remain accessible as a custom GPT. OpenAI views this release as a pivotal advancement in AI image generation, with plans for future updates to enhance editing precision, visual richness, and language support.