ChatGPT's new image generator can now 'think' before creating visuals
What's the story
OpenAI has unveiled an upgraded version of its AI-powered image generator, dubbed ChatGPT Images 2.0. The new model comes with "thinking capabilities," enabling it to scour the web for information and create multiple images from a single prompt. The update was announced by OpenAI on Tuesday and is powered by the company's latest GPT Image 2 model.
Enhanced capabilities
ChatGPT Images 2.0 can now pull information from web
The new thinking capabilities of ChatGPT Images 2.0 are available to ChatGPT Plus, Pro, Business, and Enterprise subscribers. When activated, these features allow the chatbot's image generator to pull information from the web. It can also create visual explainers based on uploaded files and "reason through the structure of the image before generating." This makes it more sophisticated at following instructions and preserving details as per user preferences.
Multi-image generation
It can also produce multiple images in one go
With the new thinking model, ChatGPT Images 2.0 can generate up to eight images at once while keeping the same characters, objects, and styles in each scene. This feature is particularly useful for creating manga pages, a series of social graphics, or design plans for every room in a house. The update also improves the model's ability to capture defining characteristics of photos and create pixel art, manga, cinematic stills among other types of images.
Language support
The update also improves image resolution and multilingual support
The new model can generate images with a resolution of up to 2K and in more aspect ratios, from wider formats like 3:1 to taller ones like 1:3. It has also improved its performance in generating text in Japanese, Korean, Chinese, Hindi, and Bengali. This makes it a more versatile tool for users across different regions and languages.