Mobile-O: Unified Multimodal Understanding and Generation on Mobile Device
2026-02-23 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors developed Mobile-O, a small and efficient model that can understand and create images using both pictures and words, all on a mobile device. It uses a special core called the Mobile Conditioning Projector to combine language and vision features without needing much computing power. Trained on a relatively small dataset, Mobile-O performs as well or better than bigger models, but runs much faster and works entirely on devices like iPhones in about 3 seconds per image. This work shows it’s possible to have smart visual and language capabilities without relying on cloud servers. The authors have also shared their code and app publicly for others to use and improve.
multimodal modelsvision-languagediffusion modeldepthwise-separable convolutionscross-modal conditioningedge devicesvisual understandingimage generationreal-time inferencemobile AI
Authors
Abdelrahman Shaker, Ahmed Heakl, Jaseel Muhammad, Ritesh Thawkar, Omkar Thawakar, Senmao Li, Hisham Cholakkal, Ian Reid, Eric P. Xing, Salman Khan, Fahad Shahbaz Khan
Abstract
Unified multimodal models can both understand and generate visual content within a single architecture. Existing models, however, remain data-hungry and too heavy for deployment on edge devices. We present Mobile-O, a compact vision-language-diffusion model that brings unified multimodal intelligence to a mobile device. Its core module, the Mobile Conditioning Projector (MCP), fuses vision-language features with a diffusion generator using depthwise-separable convolutions and layerwise alignment. This design enables efficient cross-modal conditioning with minimal computational cost. Trained on only a few million samples and post-trained in a novel quadruplet format (generation prompt, image, question, answer), Mobile-O jointly enhances both visual understanding and generation capabilities. Despite its efficiency, Mobile-O attains competitive or superior performance compared to other unified models, achieving 74% on GenEval and outperforming Show-O and JanusFlow by 5% and 11%, while running 6x and 11x faster, respectively. For visual understanding, Mobile-O surpasses them by 15.3% and 5.1% averaged across seven benchmarks. Running in only ~3s per 512x512 image on an iPhone, Mobile-O establishes the first practical framework for real-time unified multimodal understanding and generation on edge devices. We hope Mobile-O will ease future research in real-time unified multimodal intelligence running entirely on-device with no cloud dependency. Our code, models, datasets, and mobile application are publicly available at https://amshaker.github.io/Mobile-O/