Mathematical Proficiency on AGIEval MATH Level-5
64.45AccuracyHierarchical MAS + MASPO
Evaluation Results
| Method | Links | |
|---|---|---|
| Hierarchical MAS + MASPOPrompt Opt.=true, Joint Opt.=true2026.05 | 64.45 | |
| Hierarchical MAS + SPOPrompt Opt.=true, Joint Opt.=false2026.05 | 63.41 | |
| Sequential MAS + MASPOPrompt Opt.=true, Joint Opt.=true2026.05 | 61.98 | |
| Hierarchical MAS + TPEPrompt Opt.=true, Joint Opt.=false2026.05 | 60.68 | |
| Sequential MAS + SPOPrompt Opt.=true, Joint Opt.=false2026.05 | 60.13 | |
| AgentDropoutPrompt Opt.=false, Joint Opt.=false2026.05 | 59.77 | |
| Sequential MASPrompt Opt.=false, Joint Opt.=false2026.05 | 59.38 | |
| Hierarchical MASPrompt Opt.=false, Joint Opt.=false2026.05 | 59.38 | |
| Sequential MAS + TPEPrompt Opt.=true, Joint Opt.=false2026.05 | 58.73 | |
| SC (CoT)Prompt Opt.=false, Joint Opt.=false2026.05 | 56.64 | |
| Self-RefinePrompt Opt.=false, Joint Opt.=false2026.05 | 56.52 | |
| VanillaPrompt Opt.=false, Joint Opt.=false2026.05 | 55.86 | |
| CoTPrompt Opt.=false, Joint Opt.=false2026.05 | 54.69 |