Scientific Foundation Model Collaboration on EywaBench Economy
80.48Utility ScoreEywaAgent
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| EywaAgentSetting=Single-Agent2026.04 | 80.48 | 19.71 | 2,791 | |
| EywaMASSetting=Multi-Agent2026.04 | 79.59 | 68.4 | 11,044 | |
| EywaOrchestraSetting=Dynamic Orchestration2026.04 | 78.3 | 49.13 | 7,117 | |
| Debate MASSetting=Multi-Agent2026.04 | 77.29 | 92.72 | 14,694 | |
| Single-LLM-AgentSetting=Single-Agent2026.04 | 76.89 | 22.67 | 4,097 | |
| Refine MASSetting=Multi-Agent2026.04 | 75.23 | 62.33 | 8,924 | |
| MoASetting=Multi-Agent2026.04 | 74.99 | 57.35 | 16,114 | |
| X-MASSetting=Multi-Agent2026.04 | 69.23 | 75.5 | 16,499 |