List your company

General Reasoning on ARC (Pass@1)

96.2Pass@1Qwen3
87.3689.65591.9594.245Oct 30, 2025

Evaluation results

MethodLinks
Qwen3Base Model=8B, Training Method=Base2025.1096.2
GRPOBase Model=8B, Training Method=GRPO2025.1095.8
GRPOExtraRolloutsBase Model=8B, Training Method=GRPOExtraRollouts2025.1095.7
ICPO†Base Model=8B, Training Method=ICPO†2025.1095.6
ICPOBase Model=8B, Training Method=ICPO2025.1095.5
GRPOExpertDomainBase Model=8B, Training Method=GRPOExpertDomain2025.1092.2
GRPOExtraRolloutsBase Model=1.7B, Training Method=GRPOExtraRollouts2025.1088.9
GRPOExpertDomainBase Model=1.7B, Training Method=GRPOExpertDomain2025.1088.8
Qwen3Base Model=1.7B, Training Method=Base2025.1088.3
GRPOBase Model=1.7B, Training Method=GRPO2025.1088.3
ICPOBase Model=1.7B, Training Method=ICPO2025.1088.1
ICPO†Base Model=1.7B, Training Method=ICPO†2025.1087.7