Nebius Token Factory
Mini toolsManzano
By Apple
Manzano couples a hybrid image tokenizer with a unified autoregressive LLM and diffusion decoder so the same model can both understand and generate images. A single vision encoder feeds two light adapters that output continuous embeddings for image-to-text tasks and discrete tokens for image generation in a shared semantic space. Trained on both understanding and generation data, Manzano reduces the usual trade off between these abilities and achieves state-of-the-art results among unified models, competitive with specialist systems, especially on text rich benchmarks.
New
Multimodal
Gen 3
Overview
Manzano is Apple’s unified multimodal model that shares a hybrid vision tokenizer for both image understanding and text-to-image generation, using one autoregressive LLM plus a diffusion decoder to reach state-of-the-art unified performance.
About Apple
Industry:
Technology, Information and Media
Company Size:
12000
Location:
Cupertino, California, US
Website:
apple.com
Tools using Manzano
No tools found for this model yet.
Last updated: January 21, 2026
