List your company
ResearchFormerly WizWand

Mathematical Reasoning on AIME 24 (Acc, Tok)

77.9AccuracyOriginal
50.44457.57264.771.828May 9, 2026
Updated 5mo ago

Evaluation results

MethodLinks
OriginalModel Scale=Qwen3-4B-Thinking-2507, # Data=–2026.0577.919,095
SFT-S1.1Model Scale=DeepSeek-R1-Distill-Qwen-32B, # Data=1K2026.0577.311,383
Budget PromptModel Scale=Qwen3-4B-Thinking-2507, # Data=–2026.0577.118,147
HINT TUNINGModel Scale=Qwen3-4B-Thinking-2507, # Data=1K2026.0573.810,233
SFT-S1.1Model Scale=DeepSeek-R1-Distill-Qwen-14B, # Data=1K2026.0569.812,483
OriginalModel Scale=DeepSeek-R1-Distill-Qwen-32B, # Data=–2026.056911,061
HINT TUNINGModel Scale=DeepSeek-R1-Distill-Qwen-32B, # Data=1K2026.0568.88,173
OriginalModel Scale=DeepSeek-R1-Distill-Qwen-14B, # Data=–2026.0568.211,523
Budget PromptModel Scale=DeepSeek-R1-Distill-Qwen-32B, # Data=–2026.0567.79,401
SFT-S1.1Model Scale=Qwen3-4B-Thinking-2507, # Data=1K2026.0567.110,929
Budget PromptModel Scale=DeepSeek-R1-Distill-Qwen-14B, # Data=–2026.056610,513
HINT TUNINGModel Scale=DeepSeek-R1-Distill-Qwen-14B, # Data=1K2026.0563.37,963
SFT-S1.1Model Scale=DeepSeek-R1-Distill-Qwen-7B, # Data=1K2026.0557.713,414
AdaptThinkModel Scale=DeepSeek-R1-Distill-Qwen-7B, # Data=40K†2026.055510,826
VeriThinkerModel Scale=DeepSeek-R1-Distill-Qwen-7B, # Data=340K2026.0554.610,031
OriginalModel Scale=DeepSeek-R1-Distill-Qwen-7B, # Data=–2026.0553.812,959
EfficientRModel Scale=DeepSeek-R1-Distill-Qwen-7B, # Data=3.2K†, Alpha=0.42026.0553.510,336
HINT TUNINGModel Scale=DeepSeek-R1-Distill-Qwen-7B, # Data=1K2026.0552.59,899
Budget PromptModel Scale=DeepSeek-R1-Distill-Qwen-7B, # Data=–2026.0551.712,517
EfficientRModel Scale=DeepSeek-R1-Distill-Qwen-7B, # Data=3.2K†, Alpha=0.12026.0551.511,550