List your company
ResearchFormerly WizWand

Mathematical Reasoning on HMMT 26 (Accuracy)

29.4Accuracyp1
22.702424.441226.1827.9188Apr 9, 2026
Updated 6mo ago

Evaluation results

MethodLinks
p1Backbone=Qwen3-4B-Instruct-2507, S=[1, 23], M=16, Training Reward Imp.=0.212026.0429.4
p1Backbone=Qwen3-4B-Instruct-2507, S=[17, 27], M=16, Training Reward Imp.=0.092026.0428.27
p1Backbone=Qwen3-4B-Instruct-2507, S=[4, 5, 17, 20], M=8, Training Reward Imp.=0.042026.0428.13
p1Backbone=Qwen3-4B-Instruct-2507, S=[23], M=32, Training Reward Imp.=0.072026.0427.98
baseBackbone=Qwen3-4B-Instruct-2507, S=/, M=/, Training Reward Imp.=/2026.0427.89
GEPABackbone=Qwen3-4B-Instruct-2507, S=[1,23]/28, M=/, Training Reward Imp.=/2026.0427.46
GEPABackbone=Qwen3-4B-Instruct-2507, S=28/[1,23], M=/, Training Reward Imp.=/2026.0427.32
GEPABackbone=Qwen3-4B-Instruct-2507, S=15/15, M=/, Training Reward Imp.=/2026.0427.08
RLBackbone=Qwen3-4B-Instruct-2507, S=all 30, M=1, Training Reward Imp.=02026.0427.04
p1Backbone=Qwen3-4B-Instruct-2507, S=[25], M=32, Training Reward Imp.=0.072026.0426.09
baseBackbone=Qwen3-1.7B, S=/, M=/, Training Reward Imp.=/2026.0423.96
GEPABackbone=Qwen3-1.7B, S=15/15, M=/, Training Reward Imp.=/2026.0423.77
GEPABackbone=Qwen3-1.7B, S=[5,26]/28, M=/, Training Reward Imp.=/2026.0423.72
GEPABackbone=Qwen3-1.7B, S=28/[5,26], M=/, Training Reward Imp.=/2026.0423.58
p1Backbone=Qwen3-1.7B, S=[10, 11], M=16, Training Reward Imp.=0.142026.0423.58
p1Backbone=Qwen3-1.7B, S=[5, 26], M=16, Training Reward Imp.=0.132026.0423.48
RLBackbone=Qwen3-1.7B, S=all 30, M=1, Training Reward Imp.=02026.0423.44
p1Backbone=Qwen3-1.7B, S=[5, 14, 20, 26], M=8, Training Reward Imp.=0.052026.0423.34
p1Backbone=Qwen3-1.7B, S=[5], M=32, Training Reward Imp.=0.252026.0423.2
p1Backbone=Qwen3-1.7B, S=[26], M=32, Training Reward Imp.=0.292026.0422.96