List your company

Mathematical Reasoning on MATH (Solve Rate %)

58.2Solve Rate (MATH)AGENTCO-OP
34.48840.64446.852.956May 19, 2026

Evaluation results

MethodLinks
AGENTCO-OPBackbone=GPT-4o-mini2026.0558.2
AFlow*Backbone=Mixed2026.0556.2
AFlowBackbone=GPT-4o-mini2026.0553.1
MultiPersonaBackbone=GPT-4o-mini2026.0550.8
CoT SCBackbone=GPT-4o-mini, Prompting Setting=5-shot2026.0550.4
MedPromptBackbone=GPT-4o-mini2026.0550
LLM-DebateBackbone=GPT-4o-mini2026.0550
CoTBackbone=GPT-4o-mini2026.0548.8
IOBackbone=GPT-4o-mini2026.0548.6
Self RefineBackbone=GPT-4o-mini2026.0546.1
ReConcileBackbone=GPT-4o-mini2026.0544.1
ADASBackbone=GPT-4o-mini2026.0535.4