
Luma AI releases the Uni-1 model, reconstructing the paradigm of image generation technology

Luma AI released the Uni-1 image generation model on March 25, which adopts a unified intelligent architecture that combines understanding and generation processes. In the RISEBench test, it surpassed Google Nano Banana 2 and OpenAI GPT Image 1.5, particularly excelling in spatial and logical reasoning. Uni-1 reduces costs by 10% to 30% in high-resolution scene generation. The model has been utilized by companies such as Publicis Groupe and Adidas, shortening the time for advertising campaigns from several months to 40 hours. Uni-1 supports self-assessment and multi-round editing, and is now available for trial on the official website, with the API gradually opening up to developers
Due to copyright restrictions, please log in to view.
Thank you for supporting legitimate content.

