List your company

Mathematical Reasoning on CMIMC25

30.3Pass@1SFT
1.2848.81716.3523.883Mar 17, 2026

Evaluation results

MethodLinks
SFTModel Backbone=Qwen2.5-7B-Math, RLVR Stage=After RLVR (†)2026.0330.367.5
OXAFullModel Backbone=Qwen2.5-7B-Math, RLVR Stage=After RLVR (†)2026.0328.460
OXAMLEModel Backbone=Qwen2.5-7B-Math, RLVR Stage=After RLVR (†)2026.0328.365
SFTModel Backbone=Qwen2.5-7B-Math, RLVR Stage=Before RLVR2026.0326.877.5
OXAMLEModel Backbone=Qwen2.5-7B-Math, RLVR Stage=Before RLVR2026.0326.675
OXAFullModel Backbone=Qwen2.5-7B-Math, RLVR Stage=Before RLVR2026.0324.880
SFTLPModel Backbone=Qwen2.5-7B-Math, RLVR Stage=After RLVR (†)2026.0323.355
SFTLPModel Backbone=Qwen2.5-7B-Math, RLVR Stage=Before RLVR2026.0320.560
OXAFullModel Backbone=Qwen2.5-1.5B-Math, RLVR Stage=After RLVR (†)2026.0317.457.5
OXAMLEModel Backbone=Qwen2.5-1.5B-Math, RLVR Stage=After RLVR (†)2026.0316.460
OXAFullModel Backbone=Qwen2.5-1.5B-Math, RLVR Stage=Before RLVR2026.0315.962.5
OXAMLEModel Backbone=Qwen2.5-1.5B-Math, RLVR Stage=Before RLVR2026.0314.957.5
SFTModel Backbone=Qwen2.5-1.5B-Math, RLVR Stage=After RLVR (†)2026.0311.945
SFTModel Backbone=Qwen2.5-1.5B-Math, RLVR Stage=Before RLVR2026.031152.5
SFTLPModel Backbone=Qwen2.5-1.5B-Math, RLVR Stage=After RLVR (†)2026.0310.745
SFTLPModel Backbone=Qwen2.5-1.5B-Math, RLVR Stage=Before RLVR2026.039.850
BaseModel Backbone=Qwen2.5-1.5B-Math, RLVR Stage=Before RLVR2026.032.435
BaseModel Backbone=Qwen2.5-7B-Math, RLVR Stage=Before RLVR2026.032.435