| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Multi-shot video generation | GroundBench | ARC Score0.426 | 10 | |
| Multi-shot Video Generation | GroundBench (40 randomly sampled scripts) | Identity Score4.18 | 5 | |
| Long-form Question Answering | GroundBench (test) | Faithfulness (Full)87.5 | 4 |