it's the same U1-8B-MoT base with an extended MT (multi-task) training phase focused on structured visual output. the benchmark jumps are significant: IGenBench I-ACC (infographic accuracy) : 4.2👉17.0 (4x) Chart Understanding: 51.3👉69.5Text Rendering: 39.8👉46.6Overall Aesthetic: 53.8👉53.3 Repo: https://github.com/OpenSenseNova/SenseNova-U1github (infographic model docs): https://github.com/OpenSenseNova/SenseNova-U1/blob/main/docs/u1\_infographic\_model.md submitted by /u...