OpenAI Launches ChatGPT Images 2.0 Model, Enhancing Complex Visual Task Processing

On April 22, OpenAI launched the ChatGPT Images 2.0 model, significantly enhancing the ability to handle complex visual tasks, with upgrades in instruction understanding, object placement and relationship expression, as well as high-density text rendering. This model supports multilingual text generation, accurately presenting non-English content in images and improving overall semantic coherence. In terms of generation capabilities, ChatGPT Images 2.0 allows for finer detail control, including small fonts, icons, UI elements, and complex compositions, with a maximum output resolution of 2K. Additionally, it has further strengthened style representation and realism, enabling stable generation of photo-realistic images, cinematic styles, pixel art, and comics, making it suitable for scenarios such as game development, storyboard design, and marketing material production. It possesses end-to-end task processing capabilities, completing the entire workflow from copy generation to design composition. ChatGPT Images 2.0 is now available to all ChatGPT and Codex users, with the image functionality featuring ‘thinking capabilities’ accessible to Plus, Pro, and Business users (Enterprise support coming soon). The underlying model, gpt-image-2, is also available for API integration.