List your company
ResearchFormerly WizWand

Mathematical Reasoning on AIME24 (ACC, A.Tok, epsilon^3)

56.67Accuracy (ACC)VPG-EA
18.533228.434138.33548.2359May 10, 2026
Updated 5mo ago

Evaluation results

MethodLinks
VPG-EABackbone=DeepSeek-R1-Distill-Qwen-7B, Same data and RL paradigm=true2026.0556.676,152.830.52
Original ModelBackbone=DeepSeek-R1-Distill-Qwen-7B2026.0553.338,208.730.35
R1-VeriThinkerBackbone=DeepSeek-R1-Distill-Qwen-7B2026.0553.336,127.470.46
JETBackbone=DeepSeek-R1-Distill-Qwen-7B2026.0543.33——
AdaptThinkBackbone=DeepSeek-R1-Distill-Qwen-7B, Same data and RL paradigm=true2026.0543.338,044.070.23
LaserBackbone=DeepSeek-R1-Distill-Qwen-7B, Same data and RL paradigm=true2026.05404,877.270.33
AutoThinkBackbone=DeepSeek-R1-Distill-Qwen-7B, Same data and RL paradigm=true2026.05406,125.50.26
VPG-EABackbone=DeepSeek-R1-Distill-Qwen-1.5B, Same data and RL paradigm=true2026.0533.336,659.730.17
JETBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.0526.676,159.30.12
LaserBackbone=DeepSeek-R1-Distill-Qwen-1.5B, Same data and RL paradigm=true2026.0526.678,457.90.08
AutoThinkBackbone=DeepSeek-R1-Distill-Qwen-1.5B, Same data and RL paradigm=true2026.0526.676,325.20.11
AdaptThinkBackbone=DeepSeek-R1-Distill-Qwen-1.5B, Same data and RL paradigm=true2026.0523.337,596.630.07
Original ModelBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.052011,524.670.03
ThinklessBackbone=DeepSeek-R1-Distill-Qwen-1.5B2026.05207,004.40.06