List your company

Sibling-model judgment on Math AMC12 + AIME24 25 + HMMTFeb2025

0.97AUROCGnosis-SelfJudge
0.85560.88530.9150.9447Dec 23, 2025

Evaluation results

MethodLinks
Gnosis-SelfJudgeJudged Model=Qwen3 8B-Hybrid2025.120.970.970.08
Gnosis-Qwen1.7B as RMJudged Model=Qwen3 8B-Hybrid2025.120.970.990.07
Gnosis-SelfJudgeJudged Model=Qwen3 4B-Thinking2025.120.960.980.05
Gnosis-SelfJudgeJudged Model=Qwen3 1.7B-Hybrid2025.120.950.950.09
Gnosis-Qwen1.7B as RMJudged Model=Qwen3 4B-Thinking2025.120.930.970.18
SkyworkRM-Qwen3-8BJudged Model=Qwen3 1.7B-Hybrid2025.120.90.920.14
SkyworkRM-Qwen3-8BJudged Model=Qwen3 4B-Thinking2025.120.890.940.22
SkyworkRM-Qwen3-8BJudged Model=Qwen3 8B-Hybrid2025.120.860.950.23