List your company

Reasoning and Knowledge on Out-of-Distribution Benchmarks Summary

75.3Average ScoreICPO†
53.4659.1364.870.47Oct 30, 2025

Evaluation results

MethodLinks
ICPO†Base Model=8B, Training Method=ICPO†2025.1075.32.4
ICPOBase Model=8B, Training Method=ICPO2025.1074.31.4
GRPOExtraRolloutsBase Model=8B, Training Method=GRPOExtraRollouts2025.1073.91
GRPOBase Model=8B, Training Method=GRPO2025.1072.9—
GRPOExpertDomainBase Model=8B, Training Method=GRPOExpertDomain2025.1071.1-1.8
Qwen3Base Model=8B, Training Method=Base2025.1067.4—
ICPO†Base Model=1.7B, Training Method=ICPO†2025.1059.70.7
GRPOBase Model=1.7B, Training Method=GRPO2025.1059—
GRPOExpertDomainBase Model=1.7B, Training Method=GRPOExpertDomain2025.10590
GRPOExtraRolloutsBase Model=1.7B, Training Method=GRPOExtraRollouts2025.1058.2-0.8
ICPOBase Model=1.7B, Training Method=ICPO2025.1057.1-1.9
Qwen3Base Model=1.7B, Training Method=Base2025.1054.3—