| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Document Visual Question Answering | NiM-Benchmark | Score (Menus)0.63 | 24 | |
| Reinforcement Learning | Nim Gymnasium | Mean Best Reward1 | 5 | |
| Reinforcement Learning | Nim | Mean Best Reward1 | 4 | |
| Reinforcement Learning | Nim | Mean Reward0.61 | 4 | |
| P/N position classification | NIM (N=20, k=4) single-frame | Shootouts (avg)2.87 | 4 |