List your company
ResearchFormerly WizWand

Mathematics on MATH 500

97.3Pass@1DeepSeek-R1
46.038459.346772.65585.9633Jan 22, 2025Mar 30, 2025Jun 5, 2025Aug 12, 2025Oct 18, 2025Dec 24, 2025Mar 2, 2026
Updated 5mo ago

Evaluation results

MethodLinks
DeepSeek-R1Architecture=MoE, Activated Params=37B, Total Params=671B2025.0197.3
MiMo-RL 7B-R-TAPModel Backbone=7B2026.0397.3
OpenAI-o1-12172025.0196.4
MiMo 7B-RLModel Backbone=7B2026.0395.8
R1-Distill-Qwen-14BModel Backbone=Qwen-14B2026.0393.9
LearnedBackbone Model=DS-R1-Qwen-7B, Cmax=3, Bmax=92162025.0692.89
R1-Distill-Qwen-7BModel Backbone=Qwen-7B2026.0392.8
WaitBackbone Model=DS-R1-Qwen-7B, Cmax=3, Bmax=92162025.0692.17
HmmBackbone Model=DS-R1-Qwen-7B, Cmax=3, Bmax=92162025.0691.74
INVERSE-EVIDENCEModel=QWEN3-4B2026.0391.23
RANDOMModel=QWEN3-4B2026.0391.22
INSIGHTModel=QWEN3-4B2026.0391.22
MoPPSModel=QWEN3-4B2026.0390.82
DSModel=R1-DISTILL-QWEN-7B2026.0390.82
QwQ-32B PreviewModel Backbone=QwQ-32B2026.0390.6
RANDOMModel=R1-DISTILL-QWEN-7B2026.0390.5
INSIGHTModel=R1-DISTILL-QWEN-7B2026.0390.45
EXPECTED-DIFFICULTYModel=R1-DISTILL-QWEN-7B2026.0390.43
MoPPSModel=R1-DISTILL-QWEN-7B2026.0390.4
INVERSE-EVIDENCEModel=R1-DISTILL-QWEN-7B2026.0390.4
DeepSeek-V3Architecture=MoE, Activated Params=37B, Total Params=671B2025.0190.2
LearnedBackbone Model=DS-R1-Llama-8B, Cmax=3, Bmax=92162025.0690.04
OpenAI-o1-mini2025.0190
OpenAI o1-miniModel Backbone=o1-mini2026.0390
AlternativelyBackbone Model=DS-R1-Qwen-7B, Cmax=3, Bmax=92162025.0689.84
Qwen-3-30B-A3BOpenness=Open-weights, Regional Origin=Non-European, Post-training=Instruction-tuned2026.0289.7
Baseline (w/o BF)Backbone Model=DS-R1-Qwen-7B, Cmax=3, Bmax=92162025.0689.39
CritiqueBackbone Model=DS-R1-Qwen-7B, Cmax=3, Bmax=92162025.0689.06
Gemma-3-27BOpenness=Open-weights, Regional Origin=Non-European, Post-training=Instruction-tuned2026.0288.5
Qwen-3-14BOpenness=Open-weights, Regional Origin=Non-European, Post-training=Instruction-tuned2026.0286.9
HmmBackbone Model=DS-R1-Llama-8B, Cmax=3, Bmax=92162025.0686.21
WaitBackbone Model=DS-R1-Llama-8B, Cmax=3, Bmax=92162025.0686.01
Scaf-GRPOBase Model=DeepSeek-R1-Distill-Qwen-1.5B2025.1085.8
OLMo-3.1-32BOpenness=Fully-open, Regional Origin=Non-European, Post-training=Instruction-tuned2026.0285.7
Qwen-3-32BOpenness=Open-weights, Regional Origin=Non-European, Post-training=Instruction-tuned2026.0285.7
Gemma-3-12BOpenness=Open-weights, Regional Origin=Non-European, Post-training=Instruction-tuned2026.0285.3
GT-RewardBase Model=Qwen3-8B-Base2025.1184.72
OLMo-3-7BOpenness=Fully-open, Regional Origin=Non-European, Post-training=Instruction-tuned2026.0284.2
DeepSeek-R1-Distill-Qwen-1.5BBase Model=DeepSeek-R1-Distill-Qwen-1.5B2025.1083.9
Vanilla GRPOBase Model=DeepSeek-R1-Distill-Qwen-1.5B2025.1083.9
GT-RewardBase Model=Qwen3-4B-Base2025.1183.4
LearnedBackbone Model=DS-R1-Qwen-1.5B, Cmax=3, Bmax=92162025.0682.51
AlternativelyBackbone Model=DS-R1-Llama-8B, Cmax=3, Bmax=92162025.0681.86
Mistral-3.2-24BOpenness=Open-weights, Regional Origin=European, Post-training=Instruction-tuned2026.0281.5
WaitBackbone Model=DS-R1-Qwen-1.5B, Cmax=3, Bmax=92162025.0681.22
HmmBackbone Model=DS-R1-Qwen-1.5B, Cmax=3, Bmax=92162025.0680.84
AlternativelyBackbone Model=DS-R1-Qwen-1.5B, Cmax=3, Bmax=92162025.0680.33
CritiqueBackbone Model=DS-R1-Llama-8B, Cmax=3, Bmax=92162025.0680.27
Scaf-GRPOBase Model=Qwen2.5-Math-7B2025.1080
Self-HarmonyBase Model=Qwen3-8B-Base2025.1180
Baseline (w/o BF)Backbone Model=DS-R1-Qwen-1.5B, Cmax=3, Bmax=92162025.0679.88
Baseline (w/o BF)Backbone Model=DS-R1-Llama-8B, Cmax=3, Bmax=92162025.0679.56
CritiqueBackbone Model=DS-R1-Qwen-1.5B, Cmax=3, Bmax=92162025.0679.49
Majority-VotingBase Model=Qwen3-8B-Base2025.1178.99
Co-RewardBase Model=Qwen3-8B-Base2025.1178.92
IntuitorBase Model=Qwen3-8B-Base2025.1178.77
Self-HarmonyBase Model=Qwen3-4B-Base2025.1178.5
Claude-3.5-Sonnet-10222025.0178.3
Claude 3.5 Sonnet-1022Model Backbone=Claude 3.5 Sonnet2026.0378.3
Oat-ZeroBase Model=Qwen2.5-Math-7B2025.1078
Scaf-GRPOBase Model=Qwen2.5-7B2025.1077.8
Vanilla GRPOBase Model=Qwen2.5-7B2025.1077.6
RentBase Model=Qwen3-8B-Base2025.1177.26
SimpleRL-ZeroBase Model=Qwen2.5-Math-7B2025.1076.8
Co-RewardBase Model=Qwen3-4B-Base2025.1176.54
Vanilla GRPOBase Model=Qwen2.5-Math-7B2025.1075.8
Majority-VotingBase Model=Qwen3-4B-Base2025.1175.75
LUFFYBase Model=Qwen2.5-Math-7B2025.1075.2
GPT-4o-05132025.0174.6
Llama-3.3-70BOpenness=Open-weights, Regional Origin=Non-European, Post-training=Instruction-tuned2026.0274.6
RentBase Model=Qwen3-4B-Base2025.1174.6
GPT-4o 0513Model Backbone=GPT-4o2026.0374.6
GT-RewardBase Model=Llama-3.1-8B-Instruct2025.1173.6
Scaf-GRPOBase Model=Qwen2.5-Math-1.5B2025.1073.4
IntuitorBase Model=Qwen3-4B-Base2025.1172.35
Vanilla GRPOBase Model=Qwen2.5-Math-1.5B2025.1072.2
GT-RewardBase Model=Qwen3-1.7B-Base2025.1171.8
INVERSE-EVIDENCEModel=QWEN3-0.6B2026.0371.8
INSIGHTModel=QWEN3-0.6B2026.0371.75
RANDOMModel=QWEN3-0.6B2026.0371.3
MoPPSModel=QWEN3-0.6B2026.0370.6
GT-RewardBase Model=Llama-3.2-3B-Instruct2025.1169.8
Self-HarmonyBase Model=Qwen3-1.7B-Base2025.1169.6
Before RLBase Model=Qwen3-8B-Base2025.1166.8
Co-RewardBase Model=Qwen3-1.7B-Base2025.1164.67
Majority-VotingBase Model=Qwen3-1.7B-Base2025.1164.64
Qwen2.5-7BBase Model=Qwen2.5-7B2025.1061.8
RentBase Model=Qwen3-1.7B-Base2025.1161.08
Before RLBase Model=Qwen3-4B-Base2025.1160.2
Scaf-GRPOBase Model=Llama-3.2-3B-Instruct2025.1056.2
Self-HarmonyBase Model=Llama-3.2-3B-Instruct2025.1155.4
Co-RewardBase Model=Llama-3.2-3B-Instruct2025.1155.22
EuroLLM-22BOpenness=Fully-open, Regional Origin=European, Post-training=Instruction-tuned2026.0254.5
Qwen2.5-Math-7BBase Model=Qwen2.5-Math-7B2025.1053.6
Vanilla GRPOBase Model=Llama-3.2-3B-Instruct2025.1051.8
IntuitorBase Model=Qwen3-1.7B-Base2025.1151.12
Self-HarmonyBase Model=Llama-3.1-8B-Instruct2025.1150.4
Llama-3.1-8BOpenness=Open-weights, Regional Origin=Non-European, Post-training=Instruction-tuned2026.0249.4
IntuitorBase Model=Llama-3.1-8B-Instruct2025.1148.37
Co-RewardBase Model=Llama-3.1-8B-Instruct2025.1148.01