Manzano

MANZANO is a unified multimodal model developed by Apple, designed to understand and generate images within a single architecture. Its goal is to simplify and scale the use of multimodal AI through a hybrid vision tokenizer, which combines continuous representations for analysis with discrete tokens for generation. This lets the model handle both visual interpretation tasks (image-to-text) and creation tasks (text-to-image) efficiently and in an integrated way.

‍

‍

MANZANO does not yet have a defined pricing model for the market, since it is early-stage academic research. Still, given the nature of the project and Apple's track record, it is expected to eventually be built into the ecosystem's products and services — such as the iPhone, iPad, Vision Pro, and even iCloud — adding value without necessarily being sold as a separate service.

Since it is still a research project, no commercial plans have been announced. If it follows Apple's usual pattern, MANZANO will likely be embedded in existing products, working as a premium feature within devices and services rather than being offered as a standalone tool.

MANZANO represents a milestone in multimodal model research by unifying image understanding and generation in a simple, scalable architecture. Although its impact is still confined to academia, Apple is likely to turn it into a competitive advantage within its ecosystem. For the market, the model reinforces the trend of integrated multimodality as the next step for artificial intelligence applied to consumer products.