Multi-agent aggregation on 42-task multi-agent evaluation set
92.9BT-WRdiv_mixed+judge
Evaluation results
| Method | Links | ||
|---|---|---|---|
| div_mixed+judgeRel. Cost=1.0×, Exploratory=true2026.03 | 92.9 | 88.7 | |
| div_strong+judgeRel. Cost=2.5×2026.03 | 81 | 76.8 | |
| homo_opus+judgeRel. Cost=2.0×2026.03 | 51.2 | 50 | |
| div_strong+voteRel. Cost=1.9×2026.03 | 49.6 | 42.5 | |
| div_strong+synthRel. Cost=1.9×2026.03 | 17.9 | 12.7 |