List your company
ResearchFormerly WizWand

Mathematical Reasoning on AIME 26

86.7ScoreGpt-oss-20b-high
-3.46819.94143.3566.759Mar 17, 2026Mar 29, 2026Apr 10, 2026Apr 22, 2026May 4, 2026May 16, 2026May 28, 2026
Updated 5mo ago

Evaluation results

MethodLinks
Gpt-oss-20b-highGroup=Larger, Generation Configuration=Standardized2026.0386.7
Qwen3-14BGroup=Larger, Generation Configuration=Standardized2026.0386.7
Qwen3-30B-A3BGroup=Larger, Generation Configuration=Standardized2026.0386.7
EngGPT2-16B-A3BGroup=Comparable, Generation Configuration=Standardized2026.0370
p1Backbone=Qwen3-4B-Instruct-2507, S=[1, 23], M=16, Training Reward Imp.=0.212026.0462.24
p1Backbone=Qwen3-4B-Instruct-2507, S=[17, 27], M=16, Training Reward Imp.=0.092026.0457.24
p1Backbone=Qwen3-4B-Instruct-2507, S=[4, 5, 17, 20], M=8, Training Reward Imp.=0.042026.0455.68
p1Backbone=Qwen3-4B-Instruct-2507, S=[23], M=32, Training Reward Imp.=0.072026.0455.21
RLBackbone=Qwen3-4B-Instruct-2507, S=all 30, M=1, Training Reward Imp.=02026.0454.58
baseBackbone=Qwen3-4B-Instruct-2507, S=/, M=/, Training Reward Imp.=/2026.0454.38
GEPABackbone=Qwen3-4B-Instruct-2507, S=15/15, M=/, Training Reward Imp.=/2026.0454.22
GEPABackbone=Qwen3-4B-Instruct-2507, S=28/[1,23], M=/, Training Reward Imp.=/2026.0453.91
GEPABackbone=Qwen3-4B-Instruct-2507, S=[1,23]/28, M=/, Training Reward Imp.=/2026.0453.49
p1Backbone=Qwen3-4B-Instruct-2507, S=[25], M=32, Training Reward Imp.=0.072026.0451.25
p1Backbone=Qwen3-1.7B, S=[5, 26], M=16, Training Reward Imp.=0.132026.0437.6
gemma-3-12b-itGroup=Larger, Generation Configuration=Standardized2026.0336.7
p1Backbone=Qwen3-1.7B, S=[10, 11], M=16, Training Reward Imp.=0.142026.0436.15
p1Backbone=Qwen3-1.7B, S=[5, 14, 20, 26], M=8, Training Reward Imp.=0.052026.0435.99
RLBackbone=Qwen3-1.7B, S=all 30, M=1, Training Reward Imp.=02026.0435.89
p1Backbone=Qwen3-1.7B, S=[5], M=32, Training Reward Imp.=0.252026.0435.63
baseBackbone=Qwen3-1.7B, S=/, M=/, Training Reward Imp.=/2026.0435.57
GEPABackbone=Qwen3-1.7B, S=28/[5,26], M=/, Training Reward Imp.=/2026.0435.52
GEPABackbone=Qwen3-1.7B, S=[5,26]/28, M=/, Training Reward Imp.=/2026.0435.26
GEPABackbone=Qwen3-1.7B, S=15/15, M=/, Training Reward Imp.=/2026.0435.1
p1Backbone=Qwen3-1.7B, S=[26], M=32, Training Reward Imp.=0.292026.0433.23
Entropy-Cut MHModel=Qwen2.5-Math-7B2026.0513.1
SMCModel=Qwen2.5-Math-7B2026.0511.8
TMCModel=Qwen2.5-Math-7B2026.0510.6
Entropy-Cut MHModel=Qwen3-8B-Base2026.0510.3
SMCModel=Qwen3-8B-Base2026.0510.1
Moonlight-16B-A3B-InstructGroup=Comparable, Generation Configuration=Standardized2026.0310
TMCModel=Qwen3-8B-Base2026.059.8
Uniform-Cut MHModel=Qwen2.5-Math-7B2026.059.5
Uniform-Cut MHModel=Qwen3-8B-Base2026.059.5
TMCModel=Qwen2.5-7B2026.059.4
Entropy-Cut MHModel=Qwen2.5-7B2026.059.4
Entropy-Cut MHModel=Phi-4-mini-instruct2026.058.7
SMCModel=Phi-4-mini-instruct2026.058.6
Uniform-Cut MHModel=Qwen2.5-7B2026.058.2
Uniform-Cut MHModel=Phi-4-mini-instruct2026.058
SMCModel=Qwen2.5-7B2026.057.4
TMCModel=Phi-4-mini-instruct2026.056.9
Low-TemperatureModel=Phi-4-mini-instruct2026.056.8
Low-TemperatureModel=Qwen2.5-Math-7B2026.056.4
Low-TemperatureModel=Qwen3-8B-Base2026.056.4
Low-TemperatureModel=Qwen2.5-7B2026.055.7
StandardModel=Qwen2.5-Math-7B2026.054.8
StandardModel=Phi-4-mini-instruct2026.053.5
StandardModel=Qwen3-8B-Base2026.053.2
Entropy-Cut MHModel=Phi-3.5-mini-instruct2026.053.1
TMCModel=Phi-3.5-mini-instruct2026.053
Uniform-Cut MHModel=Phi-3.5-mini-instruct2026.052.9
StandardModel=Qwen2.5-7B2026.052
SMCModel=Phi-3.5-mini-instruct2026.051.7
Low-TemperatureModel=Phi-3.5-mini-instruct2026.051.6
StandardModel=Phi-3.5-mini-instruct2026.050.6
Llama-3.1-8B-InstructGroup=Comparable, Generation Configuration=Standardized2026.030.1
gemma-2-9b-itGroup=Comparable, Generation Configuration=Standardized2026.030