List your company

Mathematical Reasoning on AIME 25 (accuracy)

94.5AccuracyAPRM
52.84863.661574.47585.2885Nov 28, 2025Dec 10, 2025Dec 23, 2025Jan 5, 2026Jan 17, 2026Jan 30, 2026Feb 12, 2026

Evaluation results

MethodLinks
APRMBackbone=GPT-OSS-120B2025.1194.5
APRMBackbone=GPT-OSS-20B2025.1193.6
Self-ConsistencyModel=DeepSeek-V3.2, API Calls=O(N)2026.0193.3
Length-based RankingModel=DeepSeek-V3.2, API Calls=O(N)2026.0193.3
Self-VerificationModel=DeepSeek-V3.2, API Calls=O(NL)2026.0193.3
UniCogModel=DeepSeek-V3.2, API Calls=O(N)2026.0193.3
LLM-JBackbone=GPT-OSS-20B2025.1192.8
APRMBackbone=Gemma-3-27B2025.1191.9
LLM-JBackbone=Gemma-3-27B2025.1191.6
StandardModel=DeepSeek-V3.2, API Calls=O(1)2026.0191.3
ReST-MCTSBackbone=GPT-OSS-20B2025.1191.3
ReST-MCTSBackbone=GPT-OSS-120B2025.1191
rStar MathBackbone=GPT-OSS-20B2025.1190.8
MajorityModel=PaCoRe-8B, Sampling Mode=Parallel, Sample Count=@5122026.0290.7
rStar MathBackbone=GPT-OSS-120B2025.1190.7
ReST-MCTSBackbone=Gemma-3-27B2025.1190.6
LLM-JBackbone=GPT-OSS-120B2025.1190.4
ToTBackbone=GPT-OSS-20B2025.1190.4
rStar MathBackbone=Gemma-3-27B2025.1189.5
ToTBackbone=GPT-OSS-120B2025.1189.2
AutoPSVBackbone=GPT-OSS-120B2025.1188
ToTBackbone=Gemma-3-27B2025.1188
Math S.Backbone=GPT-OSS-120B2025.1187.5
DeepConfModel=DeepSeek-8B, Sample Count=@5122026.0287.4
DeepConfModel=DeepSeek-8B, Sample Count=@202026.0287.4
DeepConfModel=PaCoRe-8B, Sample Count=@202026.0287.4
CoRefine TreeModel=DeepSeek-8B2026.0287.3
CoRefine TreeModel=PaCoRe-8B2026.0287.3
MajorityModel=DeepSeek-8B, Sampling Mode=Parallel, Sample Count=@202026.0286.7
CoRefineModel=DeepSeek-8B2026.0286.7
DeepConfModel=PaCoRe-8B, Sample Count=@5122026.0286.7
CoRefineModel=PaCoRe-8B2026.0286.7
MajorityModel=DeepSeek-8B, Sampling Mode=Sequential, Sample Count=@202026.0286
Verify SbS.Backbone=GPT-OSS-120B2025.1186
CoT-SCBackbone=GPT-OSS-120B2025.1185.5
AutoPSVBackbone=GPT-OSS-20B2025.1185
DeepConfModel=Qwen3-32B, Sample Count=@202026.0283.3
CoRefine TreeModel=Qwen3-32B2026.0283.3
MajorityModel=PaCoRe-8B, Sampling Mode=Parallel, Sample Count=@202026.0283.3
MajorityModel=PaCoRe-8B, Sampling Mode=Sequential, Sample Count=@202026.0283.3
CoRefineModel=Qwen3-32B2026.0282.7
MajorityModel=Qwen3-32B, Sampling Mode=Parallel, Sample Count=@202026.0282.6
MajorityModel=DeepSeek-8B, Sampling Mode=Parallel, Sample Count=@5122026.0282
MajorityModel=Qwen3-32B, Sampling Mode=Sequential, Sample Count=@202026.0282
Math S.Backbone=GPT-OSS-20B2025.1182
DeepConfModel=Qwen3-32B, Sample Count=@5122026.0281.3
Falcon-H1R# Param.=7B2026.0281.04
MajorityModel=Qwen3-32B, Sampling Mode=Parallel, Sample Count=@5122026.0280.7
Verify SbS.Backbone=GPT-OSS-20B2025.1180
AutoPSVBackbone=Gemma-3-27B2025.1180
MiniCPM-SALA# Param.=9B2026.0278.33
Math S.Backbone=Gemma-3-27B2025.1178
DeepSeek-8BStrategy=Pass, Sample Count=@12026.0277.4
Ministral-3-R# Param.=8B2026.0275
Verify SbS.Backbone=Gemma-3-27B2025.1175
DyLANMAS-Arch=Adaptive, ProVe Level=Iteration, Verifier Type=RM2026.0273.33
PaCoRe-8BStrategy=Pass, Sample Count=@12026.0273.3
DyLANMAS-Arch=Adaptive, ProVe Level=Agent, Verifier Type=RM2026.0272.22
MiniCPM-4.1# Param.=8B2026.0272.08
DyLANMAS-Arch=Adaptive, ProVe Level=Iteration, Verifier Type=Judge2026.0271.11
Qwen3-32BStrategy=Pass, Sample Count=@12026.0270.7
DyLANMAS-Arch=Adaptive, ProVe Level=Agent, Verifier Type=Judge2026.0270
DemyAgent-4BStarting Checkpoint=Inst, Backbone Architecture Family=Qwen3-4B2026.0270
CoT-SCBackbone=GPT-OSS-20B2025.1169.9
CoT-SCBackbone=Gemma-3-27B2025.1169.9
ELPO-4BStarting Checkpoint=Inst, Backbone Architecture Family=Qwen3-4B2026.0269.4
MaASMAS-Arch=Adaptive, ProVe Level=Iteration, Verifier Type=Judge2026.0268.06
DyLANMAS-Arch=Adaptive, ProVe Level=Agent, Verifier Type=PRM2026.0267.78
DyLANMAS-Arch=Adaptive, ProVe Level=Iteration, Verifier Type=PRM2026.0267.77
Qwen3# Param.=8B2026.0266.67
APRMBackbone=Gemma-3-12B2025.1165.7
DyLANMAS-Arch=Adaptive, ProVe Level=No ProVe, Verifier Type=None2026.0265.55
MaASMAS-Arch=Adaptive, ProVe Level=Iteration, Verifier Type=RM2026.0263.89
MaASMAS-Arch=Adaptive, ProVe Level=Iteration, Verifier Type=PRM2026.0263.88
DebateMAS-Arch=Fixed, ProVe Level=Agent, Verifier Type=Judge2026.0262.97
ReST-MCTSBackbone=Gemma-3-12B2025.1162.6
rStar MathBackbone=Gemma-3-12B2025.1162
DebateMAS-Arch=Fixed, ProVe Level=No ProVe, Verifier Type=None2026.0261.87
ToTBackbone=Gemma-3-12B2025.1161.4
ADASMAS-Arch=Fixed, ProVe Level=Agent, Verifier Type=RM2026.0261.1
ADASMAS-Arch=Fixed, ProVe Level=Iteration, Verifier Type=Judge2026.0261.1
MaASMAS-Arch=Adaptive, ProVe Level=Agent, Verifier Type=Judge2026.0261.09
AFlowMAS-Arch=Fixed, ProVe Level=Iteration, Verifier Type=Judge2026.0260.39
ADASMAS-Arch=Fixed, ProVe Level=Iteration, Verifier Type=RM2026.0259.73
AFlowMAS-Arch=Fixed, ProVe Level=No ProVe, Verifier Type=None2026.0259.72
MaASMAS-Arch=Adaptive, ProVe Level=Agent, Verifier Type=PRM2026.0259.7
AFlowMAS-Arch=Fixed, ProVe Level=Agent, Verifier Type=Judge2026.0258.33
ADASMAS-Arch=Fixed, ProVe Level=Agent, Verifier Type=Judge2026.0258.33
MaASMAS-Arch=Adaptive, ProVe Level=No ProVe, Verifier Type=None2026.0258.33
AFlowMAS-Arch=Fixed, ProVe Level=Agent, Verifier Type=RM2026.0258.32
AutoPSVBackbone=Gemma-3-12B2025.1158
DebateMAS-Arch=Fixed, ProVe Level=Iteration, Verifier Type=RM2026.0257.78
DebateMAS-Arch=Fixed, ProVe Level=Agent, Verifier Type=PRM2026.0257.12
Math S.Backbone=Gemma-3-12B2025.1157
AFlowMAS-Arch=Fixed, ProVe Level=Agent, Verifier Type=PRM2026.0256.94
Nemotron-Nano-v2# Param.=9B2026.0256.67
LLM-JBackbone=Gemma-3-12B2025.1155.4
Verify SbS.Backbone=Gemma-3-12B2025.1155
DebateMAS-Arch=Fixed, ProVe Level=Iteration, Verifier Type=PRM2026.0254.59
DebateMAS-Arch=Fixed, ProVe Level=Agent, Verifier Type=RM2026.0254.45