OpenAI has released ChatGPT Images 2.0, an image generator — The Verge
OpenAI has begun rolling out an updated version of its visual tool, ChatGPT Images 2, which now features a revolutionary ability to "think."
This is reported by The Verge.
The model can now use web search to analyze queries more deeply, allowing it to generate image series with high accuracy and detail.
The main feature of the update is the generator’s ability to analyze the structure of the future image before starting work.
This allows the model to better follow complex instructions, preserve specific details, and reproduce text captions directly on images with significantly higher quality.
Intelligent analysis features are currently available to Plus, Pro, Business, and Enterprise subscribers.
In this mode, ChatGPT Images 2 can independently search for information online to visualize specific concepts or create explanatory diagrams based on files uploaded by the user.
In "thinking" mode, the system can generate up to eight images simultaneously while strictly adhering to a consistent style.
This opens up new possibilities for comic book authors and designers, as the model has learned to preserve the identity of characters and objects across different scenes.
At the same time, core updates have become available to a wider range of service users.
The generator has significantly improved its ability to capture key facial features in photographs and now includes specialized settings for working with pixel art, manga, and complex cinematic compositions.
The technical specifications of the updated version include support for resolutions up to 2K.
Additionally, the developers have expanded the list of available formats—users can now create images ranging from panoramic (3:1) to narrow vertical (1:3).
OpenAI has paid special attention to multilingual support for text captions within generated content. In addition to its traditionally high-quality English, the model now reproduces Japanese, Korean, and Chinese much better, as well as Hindi and Bengali.
Market experts believe that integrating web search directly into the visual content creation process is a significant step forward.
This allows AI to go beyond mere imagination and rely on real facts and up-to-date visual data when fulfilling user requests.
Bezos’s AI startup to raise $10 billion in funding round
As a reminder, Apple removed apps for tracking ICE agents under pressure from Trump.
Apple may unveil a new affordable MacBook model next week. Insiders report eight technical limitations of the model and predict its likely price for Ukrainian buyers.
