Apple Unveils New Image Generation Model: Manzano
3 day ago / Read about 0 minute
Author:小编   

Apple has unveiled a cutting-edge image model dubbed Manzano, which is adept at performing both image understanding and generation tasks concurrently. While it hasn't been made available to the general public or subjected to public demonstrations yet, Apple's research team has highlighted its prowess in handling intricate prompts, as evidenced by research papers and the inclusion of low-resolution image samples. When pitted against the outputs of the open-source model Deepseek Janus Pro, as well as commercial systems like GPT-4o and Gemini2.5 Flash Image Generation, Manzano demonstrated comparable performance to OpenAI's GPT-4o and Google's Nano Banana across three sets of demanding prompt tests.