List your company
ResearchFormerly WizWand

Mathematical Reasoning on HARDMath2

36.5AccuracyHermes@5+Skywork
3.01211.70620.429.094Nov 24, 2025
Updated 5mo ago

Evaluation results

MethodLinks
Hermes@5+SkyworkModel=DeepSeek-V3.22025.1136.5
Hermes@5+MajorityModel=DeepSeek-V3.22025.1135.1
Hermes@1Model=DeepSeek-V3.22025.1133.6
Hermes@5+MajorityModel=OpenAI o3-mini2025.1131.8
Hermes@1Model=OpenAI o3-mini2025.1131.3
Hermes@5+SkyworkModel=OpenAI o3-mini2025.1131.3
CoT@5+SkyworkModel=DeepSeek-V3.22025.1130.8
CoT@5+Safe*Model=DeepSeek-V3.22025.1130.3
CoT@5+SkyworkModel=OpenAI o3-mini2025.1129.4
CoT@5+SafeModel=DeepSeek-V3.22025.1128.9
CoT@5+ArmoRMModel=OpenAI o3-mini2025.1128.4
CoT@5+RLHFlowModel=OpenAI o3-mini2025.1128.4
CoT@5+ArmoRMModel=DeepSeek-V3.22025.1128.4
CoT@5+ShepherdModel=DeepSeek-V3.22025.1127.8
CoT@5+SafeModel=OpenAI o3-mini2025.1126.1
CoT@5+MajorityModel=DeepSeek-V3.22025.1126.1
CoT@5+ShepherdModel=OpenAI o3-mini2025.1125.6
CoT@5+Safe*Model=OpenAI o3-mini2025.1125.6
CoT@1Model=DeepSeek-V3.22025.1125.6
CoT@5+RLHFlowModel=DeepSeek-V3.22025.1124.6
CoT@1Model=OpenAI o3-mini2025.1123.2
CoT@5+MajorityModel=OpenAI o3-mini2025.1122.7
Hermes@5+SkyworkModel=Qwen3-8B2025.1110
Hermes@5+MajorityModel=Qwen3-8B2025.117.6
Hermes@1Model=Qwen3-8B2025.116.6
CoT@5+Safe*Model=Qwen3-8B2025.116.2
CoT@5+SkyworkModel=Qwen3-8B2025.115.7
CoT@5+ShepherdModel=Qwen3-8B2025.115.7
CoT@5+RLHFlowModel=Qwen3-8B2025.115.7
CoT@5+SafeModel=Qwen3-8B2025.115.7
CoT@5+ArmoRMModel=Qwen3-8B2025.115.2
CoT@5+MajorityModel=Qwen3-8B2025.114.7
CoT@1Model=Qwen3-8B2025.114.3