1 paper · 1 filter
Kai Zou, Ziqi Huang, Yuhao Dong +7
Unified multimodal models aim to jointly enable visual understanding and generation, yet current benchmarks rarely examine their true integration. Existing evaluations either treat…