Scientific Reasoning on SciBench Physics
82.4AccuracySingle Agent
Evaluation results
| Method | Links | |
|---|---|---|
| Single AgentBase Model=GPT-4o, Aligned=false, AEA Model=None2026.05 | 82.4 | |
| Weak-to-Strong (AEA)Base Model=GPT-4o, Aligned=true, AEA Model=AEA-4B2026.05 | 76.5 | |
| Self-Reflection (AEA)Base Model=GPT-4o, Aligned=true, AEA Model=GPT-4o2026.05 | 71 | |
| Weak-to-Strong (AEA)Base Model=Claude 3.7 Sonnet, Aligned=true, AEA Model=AEA-4B2026.05 | 68.4 | |
| Self-Reflection (AEA)Base Model=Claude 3.7 Sonnet, Aligned=true, AEA Model=Claude 3.7 Sonnet2026.05 | 67.6 | |
| Single AgentBase Model=GPT-4o-mini, Aligned=false, AEA Model=None2026.05 | 67.6 | |
| Multi AgentBase Model=Claude 3.7 Sonnet, Aligned=false, AEA Model=None2026.05 | 64.7 | |
| Weak-to-Strong (AEA)Base Model=Claude 3.5 Sonnet, Aligned=true, AEA Model=AEA-4B2026.05 | 61.8 | |
| Multi AgentBase Model=GPT-4o-mini, Aligned=false, AEA Model=None2026.05 | 61.8 | |
| Self-Reflection (AEA)Base Model=GPT-4o-mini, Aligned=true, AEA Model=GPT-4o-mini2026.05 | 61.8 | |
| Weak-to-Strong (AEA)Base Model=GPT-4o-mini, Aligned=true, AEA Model=AEA-4B2026.05 | 61.8 | |
| Multi AgentBase Model=GPT-4o, Aligned=false, AEA Model=None2026.05 | 61.8 | |
| Self-Reflection (AEA)Base Model=Claude 3.5 Sonnet, Aligned=true, AEA Model=Claude 3.5 Sonnet2026.05 | 60.2 | |
| Weak-to-Strong (AEA)Base Model=Claude 3 Haiku, Aligned=true, AEA Model=AEA-4B2026.05 | 58.8 | |
| Multi AgentBase Model=Claude 3.5 Sonnet, Aligned=false, AEA Model=None2026.05 | 58.8 | |
| Single AgentBase Model=Claude 3.7 Sonnet, Aligned=false, AEA Model=None2026.05 | 58.8 | |
| Self-Reflection (AEA)Base Model=Claude 3 Haiku, Aligned=true, AEA Model=Claude 3 Haiku2026.05 | 41.2 | |
| Multi AgentBase Model=Claude 3 Haiku, Aligned=false, AEA Model=None2026.05 | 36.7 | |
| Single AgentBase Model=Claude 3.5 Sonnet, Aligned=false, AEA Model=None2026.05 | 35.3 | |
| Single AgentBase Model=Claude 3 Haiku, Aligned=false, AEA Model=None2026.05 | 23.5 |