Multi-agent Reasoning on Reasoning Benchmarks Cooperative AutoGen framework (test)
83.58Overall AccuracyMARSHAL (Generalist, 8B)
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| MARSHAL (Generalist, 8B)MAS=AutoGen, Model=MARSHAL (Generalist, 8B)2025.10 | 83.58 | 94.4 | 95 | 85.04 | 70 | 95 | 90.04 | 55.56 | |
| Qwen3-8BMAS=AutoGen, Model=Qwen3-8B2025.10 | 79.68 | 88.8 | 95.91 | 83.07 | 60 | 89.19 | 89.3 | 51.52 |