Multi-agent Reasoning on Reasoning Benchmarks Competitive MAD framework (test)
0.8509Average ScoreMARSHAL (Generalist, 8B)
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| MARSHAL (Generalist, 8B)MAS=MAD, Model=MARSHAL (Generalist, 8B)2025.10 | 0.8509 | 0.964 | 0.9659 | 0.8346 | 0.8 | 0.95 | 0.907 | 0.5354 | |
| Qwen3-8BMAS=MAD, Model=Qwen3-8B2025.10 | 0.8249 | 0.95 | 0.9636 | 0.8346 | 0.7 | 0.9 | 0.8959 | 0.5303 |