List your company

LLM-as-a-Judge on SenseBench

86.53Math ScoreSenseJudge
52.00260.96669.9378.894Jun 2, 2026

Evaluation results

MethodLinks
SenseJudgeBase Model=Qwen3-14B-Instruct2026.0686.5387.9683.6992.2475.2781.0478.7275.7882.65
SenseJudgeBase Model=Llama3.1-8B-Instruct2026.0684.5879.3471.882.3973.4369.875.6972.4474.98
SenseJudgeBase Model=Qwen2.5-72B-Instruct2026.0682.389.0179.7689.8779.8282.1278.175.2381.99
SenseJudgeBase Model=Qwen2.5-14B-Instruct2026.0673.4580.972.4485.6772.8975.2476.874.2176.88
DeepSeek-V3Category=Strong Open/Closed-Source Models2026.0672.862.2766.6777.0762.6764.464.861.8766.57
Qwen3-14B-InstructCategory=SenseJudge (Baseline)2026.0671.7364.869.27861.4767.666.2756.1366.9
Skywork-Reward-Gemma2-27BCategory=Trained Judger / Reward Models2026.0670.461.666.174.1646062.758.464.7
QRM-Gemma-2-27BCategory=Trained Judger / Reward Models2026.0669.658.9363.4754.1352.5350.456.5359.7358.17
INF-ORM-Llama3.1-70BCategory=Trained Judger / Reward Models2026.0667.263.264.869.076054.1358.459.7362.07
Gpt-4oCategory=Strong Open/Closed-Source Models2026.066661.665.4772.9360.863.265.4756.463.98
Qwen-Plus-LatestCategory=Strong Open/Closed-Source Models2026.066659.661.267.255.4758.1360.459.3360.92
Auto-j-13BCategory=Trained Judger / Reward Models2026.0664.452.6755.7353.8739.3342.44836.449.1
Qwen2.5-72B-InstructCategory=SenseJudge (Baseline)2026.0663.8761.4761.8776.2760.1366.464.457.0763.93
Skywork-Critic-Llama3.1-70BCategory=Trained Judger / Reward Models2026.0663.7363.8763.7369.8759.6759.4757.3355.4761.35
Llama3.1-8B-InstructCategory=SenseJudge (Baseline)2026.0662.1358.859.737258.1358.5360.1353.660.33
Qwen2.5-14B-InstructCategory=SenseJudge (Baseline)2026.0660.2760.859.3370.1361.66264.457.261.97
DeepSeek-R1Category=Strong Open/Closed-Source Models2026.0653.3355.661.3353.4746.9350.5351.7350.6752.95