List your company
ResearchFormerly WizWand

Mathematical Reasoning on MATH500 (Mean Accuracy, Std Dev)

98.2Mean Accuracy (MATH500)Hermes@1
83.01686.95890.994.842Nov 24, 2025
Updated 5mo ago

Evaluation results

MethodLinks
Hermes@1Model=DeepSeek-V3.22025.1198.298.2
CoT@5+SkyworkModel=DeepSeek-V3.22025.1197.397.3
CoT@5+Safe*Model=DeepSeek-V3.22025.1197.397.3
Hermes@1Model=DeepSeek-V3.12025.1197.297.2
CoT@5+MajorityModel=DeepSeek-V3.22025.1197.297.2
CoT@5+SafeModel=DeepSeek-V3.22025.1197.197.1
Hermes@1Model=OpenAI o3-mini2025.119797
CoT@5+ArmoRMModel=DeepSeek-V3.22025.119797
CoT@5+MajorityModel=DeepSeek-V3.12025.1196.996.9
CoT@5+SkyworkModel=DeepSeek-V3.12025.1196.796.7
CoT@5+ShepherdModel=DeepSeek-V3.22025.1196.796.7
CoT@1Model=DeepSeek-V3.22025.1196.696.6
CoT@5+MajorityModel=OpenAI o3-mini2025.1196.596.5
CoT@5+SkyworkModel=OpenAI o3-mini2025.1196.496.4
CoT@5+Safe*Model=OpenAI o3-mini2025.1196.496.4
CoT@5+ArmoRMModel=DeepSeek-V3.12025.1196.496.4
CoT@5+ShepherdModel=DeepSeek-V3.12025.1196.496.4
CoT@5+Safe*Model=DeepSeek-V3.12025.1196.496.4
CoT@1Model=OpenAI o3-mini2025.1196.296.2
CoT@5+ArmoRMModel=OpenAI o3-mini2025.1196.196.1
CoT@5+ShepherdModel=OpenAI o3-mini2025.1196.196.1
CoT@5+RLHFlowModel=DeepSeek-V3.12025.119696
CoT@5+SafeModel=DeepSeek-V3.12025.119696
CoT@5+RLHFlowModel=OpenAI o3-mini2025.1195.895.8
CoT@1Model=DeepSeek-V3.12025.1195.695.6
CoT@5+SafeModel=OpenAI o3-mini2025.1195.595.5
CoT@5+RLHFlowModel=DeepSeek-V3.22025.1195.595.5
Hermes@1Model=Qwen3-8B2025.1191.391.3
CoT@5+SkyworkModel=Qwen3-8B2025.1190.990.9
CoT@5+ArmoRMModel=Qwen3-8B2025.1189.289.2
CoT@5+Safe*Model=Qwen3-8B2025.1189.289.2
CoT@5+SafeModel=Qwen3-8B2025.1189.189.1
CoT@5+ShepherdModel=Qwen3-8B2025.1186.986.9
CoT@5+MajorityModel=Qwen3-8B2025.1186.886.8
CoT@1Model=Qwen3-8B2025.1185.785.7
CoT@5+RLHFlowModel=Qwen3-8B2025.1183.683.6