OpenAI announced ChatGPT Image 2.5the new model is dedicated to creating images with more accurate details and better tools. Every week, users generate more than 3 billion images using the Images model. With this new update, OpenAI further improves performancereducing build time and making the process faster. Let’s look at all the details below, obviously with a series of illustrative examples.
What’s new in ChatGPT Images 2.5
First, improvements involve lighting and textures, which are now more realistic and natural, with a higher level of detail. also very important Preserves the validity of the subject present in the reference photoso no inconsistent changes are made. Latency in generating imagery has also been reduced by 50% compared to previous models, so the process will be significantly faster.
As expected, OpenAI is also launching new tools designed to provide greater control over the creative process. For example, Sketch lets you Draw directly in the app for reference Used to create images. If you want to ask ChatGPT to make changes, you can Insert comments directly on the image.
In this regard, targeted modification operations have been simplified, making it possible to obtain More accurately reflect the requested imagekeeping all other details unchanged, even the most intricate ones.
Other improvements
Images 2.5 is now more efficient Get detailed visual instructions And can handle more complex layouts, including transparent backgrounds. If you don’t have a particularly clear idea, you can Choose between different templates Then add more details based on your needs.
OpenAI announces GPT-6 Astra, an advanced model for coding and professional work
Also, if you have successfully created a particularly interesting and potentially useful image for other users, now You can also share tips. Images 2.5 is now available for ChatGPT, ChatGPT Work and Codex users, as well as the API.
