OpenAI on the 21st (local time) unveiled ChatGPT Image 2.0, a next-generation model with significantly strengthened image generation.
The service was developed based on ImageGen 2.0. It is the official release version of Duct Tape, which was evaluated in user tests on the AI evaluation platform Arena as having greatly improved text rendering issues.
ChatGPT Image 2.0 improves the text quality that had been cited as a previous weakness. Text rendering accuracy has increased across multiple languages, including Korean, Japanese, Chinese, Hindi, and Bengali, and it can now render small text.
Image resolution is supported up to 2K, and images can be generated in aspect ratios ranging from 3:1 to 1:3. It can also create up to 10 images at a time.
In addition, it can reproduce specific styles, including icons as well as comics and films. OpenAI said it can deliver more useful results than before by precisely reflecting users' detailed instructions.
Meanwhile, Thinking and Pro models, which enhance reasoning during image generation, were unveiled as well. Using them can yield more accurate results and maintain character consistency across multiple images, such as in comics.
ChatGPT Image 2.0 is available on every account, including free users. However, advanced, reasoning-based outputs are offered only to paid users on Plus, Pro, and Business.