| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| Minecraft Long-horizon Tasks | Human | Wood100 | 15 | 3mo ago | |
| LongAct Bench (detailed split) | GPT-5 | GC (Avg)59 | 7 | 19d ago | |
| Red Dead Redemption 2 (RDR2) | SPIKE | Success Rate (SR)56.4 | 6 | 15d ago | |
| Long-horizon complex tasks (test) | Success Rate100 | 3 | 1mo ago |