OpenAI has announced the launch of ChatGPT Images 2.0, a new image model that significantly enhances the capability to handle complex visual tasks.

BlockBeats News, April 22, OpenAI launched the ChatGPT Images 2.0 image model, significantly improving the capability to handle complex visual tasks. The model has seen upgrades in instruction understanding, object positioning and relation expression, and high-density text rendering. This model supports multi-language text generation, accurately rendering non-English content in images and enhancing overall semantic coherence.

In terms of generation capability, ChatGPT Images 2.0 can achieve finer detail control, including small fonts, icons, UI elements, and complex compositions, supporting up to 2K resolution output. Furthermore, it has enhanced its style representation and realism, stably generating photo-realistic images, cinematic styles, pixel art, and comics, among other visual styles, suitable for scenarios such as game development, storyboard design, and marketing material creation. With end-to-end task processing capability, it can complete the entire process from text generation to design composition.

ChatGPT Images 2.0 is now available to all ChatGPT and Codex users, with the image feature that possesses "thinking abilities" open to Plus, Pro, and Business users (Enterprise support coming soon). The underlying model gpt-image-2 has also opened API access.