List your company

Mathematical Reasoning on GSM8K (Acc, Tokens)

97AccuracyCOT
91.831293.173194.51595.8569May 22, 2025Jul 24, 2025Sep 25, 2025Nov 27, 2025Jan 29, 2026Apr 2, 2026Jun 5, 2026

Evaluation results

MethodLinks
COTModel=QwQ-32B2026.04971,561—
COTModel=QwQ-32B2026.04971,561—
EquivPrunerBackbone=Qwen2.5-14B-Instruct2025.059727,40261.71
BaselineBackbone=QwQ-32B2026.0696.81,505—
DyConBackbone=QwQ-32B2026.0696.8995—
CyclicReflexBase Model=QwQ-32B2026.0596.71,581—
PathCalBase Model=QwQ-32B2026.0596.71,430—
NoWaitBackbone=QwQ-32B2026.0696.7983—
QwQ-32B MixReasoningMode=MixReasoning, Base Model=QwQ-32B2025.1096.64880—
SATModel=QwQ-32B2026.0496.6969—
SATModel=QwQ-32B2026.0496.6969—
Qwen3-14B MixReasoningMode=MixReasoning, Base Model=Qwen3-14B2025.1096.571,137—
SATModel=Qwen3-32B2026.0496.5680—
SATModel=Qwen3-32B2026.0496.5680—
CoDBackbone=QwQ-32B2026.0696.5617—
NothinkingBackbone=QwQ-32B2026.0696.51,331—
FlashThinkBackbone=QwQ-32B2026.0696.5910—
SATModel=Qwen3-14B2026.0496.4767—
SATModel=Qwen3-14B2026.0496.4767—
TIPBase Model=QwQ-32B2026.0596.41,494—
S1Base Model=QwQ-32B2026.0596.41,830—
QwQ-32BMode=Original, Base Model=QwQ-32B2025.1096.371,745—
DEERModel=QwQ-32B2026.0496.3977—
DEERModel=QwQ-32B2026.0496.3977—
DEERBackbone=QwQ-32B2026.0696.3977—
SEALBackbone=QwQ-32B2026.0696.31,231—
BaselineBackbone=Qwen3-14B2026.0696.31,693—
DyConBackbone=Qwen3-14B2026.0696.31,166—
Qwen3-14BMode=Original, Base Model=Qwen3-14B2025.1096.251,950—
DEERModel=Qwen3-32B2026.0496.2769—
DEERModel=Qwen3-32B2026.0496.2769—
CoDBackbone=Qwen3-14B2026.0696.2670—
OriginalBase Model=QwQ-32B2026.0596.11,593—
MERABackbone=DeepSeek-R1-Distill-Qwen-14B, Zero-shot Chain-of-Thought (CoT)=true2025.0896.1835—
COTModel=Qwen3-32B2026.04961,688—
COTModel=Qwen3-32B2026.04961,688—
Vanilla MCTSBackbone=Qwen2.5-14B-Instruct2025.059644,404100
BaselineBackbone=Qwen3-4B-Thinking-25072026.0695.91,494—
Qwen3-8B MixReasoningMode=MixReasoning, Base Model=Qwen3-8B2025.1095.881,389—
COTModel=Qwen3-8B2026.0495.82,152—
COTModel=Qwen3-8B2026.0495.82,152—
NoWaitBackbone=Qwen3-14B2026.0695.81,125—
COTModel=Qwen3-14B2026.0495.71,660—
COTModel=Qwen3-14B2026.0495.71,660—
CoDBackbone=Qwen3-4B-Thinking-25072026.0695.7952—
DEERBackbone=Qwen3-4B-Thinking-25072026.0695.71,037—
DyConBackbone=Qwen3-4B-Thinking-25072026.0695.71,098—
Dynasor-CoTBackbone=Qwen3-14B2026.0695.61,483—
QwQ-32B Concise ModeMode=Concise Mode, alpha=32, Base Model=QwQ-32B2025.1095.58664—
DEERModel=Qwen3-8B2026.0495.5981—
SATModel=Qwen3-8B2026.0495.5879—
DEERModel=Qwen3-8B2026.0495.5981—
SATModel=Qwen3-8B2026.0495.5879—
Qwen3-8BMode=Original, Base Model=Qwen3-8B2025.1095.452,403—
DEERModel=Qwen3-14B2026.0495.3840—
DEERModel=Qwen3-14B2026.0495.3840—
DEERBackbone=Qwen3-14B2026.0695.3840—
SATModel=DS-Qwen-14B2026.0495.2621—
SATModel=DS-Qwen-14B2026.0495.2621—
QwQ-32B Concise ModeMode=Concise Mode, alpha=64, Base Model=QwQ-32B2025.1095.14391—
COTModel=Qwen3-4B2026.0495.12,136—
SATModel=Qwen3-4B2026.0495.1845—
COTModel=Qwen3-4B2026.0495.12,136—
SATModel=Qwen3-4B2026.0495.1845—
NothinkingBackbone=Qwen3-4B-Thinking-25072026.06951,137—
Qwen3-14B Concise ModeMode=Concise Mode, alpha=32, Base Model=Qwen3-14B2025.1095674—
COTModel=DS-Qwen-14B2026.0494.91,122—
COTModel=DS-Qwen-14B2026.0494.91,122—
NothinkingBackbone=Qwen3-14B2026.0694.9260—
ThinkpilotBackbone=Qwen3-14B2026.0694.9274—
Qwen3-14B Concise ModeMode=Concise Mode, alpha=64, Base Model=Qwen3-14B2025.1094.87372—
NoWaitBackbone=Qwen3-4B-Thinking-25072026.0694.81,070—
DASTBackbone=DeepSeek-R1-Distill-Qwen-14B, Zero-shot Chain-of-Thought (CoT)=true2025.0894.81,778—
ThinkpilotBackbone=Qwen3-4B-Thinking-25072026.0694.7878—
FCS+Ref.Backbone=DeepSeek-R1-Distill-Qwen-14B, Zero-shot Chain-of-Thought (CoT)=true2025.0894.71,141—
DEERModel=Qwen3-4B2026.0494.51,250—
DEERModel=Qwen3-4B2026.0494.51,250—
ThinkSwitcherModel=DS-Qwen-14B2026.0494.31,042—
ThinkSwitcherModel=DS-Qwen-14B2026.0494.31,042—
S1Base Model=DeepSeek-R1-Distill-Qwen-14B2026.0593.9607—
TrimRBackbone=QwQ-32B2026.0693.71,319—
MERABackbone=DeepSeek-R1-Distill-Qwen-7B, Zero-shot Chain-of-Thought (CoT)=true2025.0893.7822—
Qwen3-8B Concise ModeMode=Concise Mode, alpha=32, Base Model=Qwen3-8B2025.1093.65817—
PathCalBase Model=DeepSeek-R1-Distill-Qwen-14B2026.0593.6479—
CyclicReflexBase Model=DeepSeek-R1-Distill-Qwen-14B2026.0593.5537—
DEERBackbone=DeepSeek-R1-Distill-Qwen-14B, Zero-shot Chain-of-Thought (CoT)=true2025.0893.5975—
DEERModel=DS-Qwen-14B2026.0493.31,040—
DEERModel=DS-Qwen-14B2026.0493.31,040—
OriginalBase Model=DeepSeek-R1-Distill-Qwen-14B2026.0593.3480—
TIPBase Model=DeepSeek-R1-Distill-Qwen-14B2026.0593482—
FCS+Ref.Backbone=DeepSeek-R1-Distill-Qwen-7B, Zero-shot Chain-of-Thought (CoT)=true2025.0892.91,218—
Qwen3-8B Concise ModeMode=Concise Mode, alpha=64, Base Model=Qwen3-8B2025.1092.89534—
DASTBackbone=DeepSeek-R1-Distill-Qwen-7B, Zero-shot Chain-of-Thought (CoT)=true2025.0892.71,558—
No WaitBackbone=DeepSeek-R1-Distill-Qwen-14B, Zero-shot Chain-of-Thought (CoT)=true2025.0892.61,152—
ThinkSwitcherModel=DS-Qwen-7B2026.0492.51,389—
ThinkSwitcherModel=DS-Qwen-7B2026.0492.51,389—
O1-PrunerBackbone=DeepSeek-R1-Distill-Qwen-7B, Zero-shot Chain-of-Thought (CoT)=true2025.0892.51,052—
DeepSeek-R1-7B MixReasoningMode=MixReasoning, Base Model=DeepSeek-R1-7B2025.1092.3765—
OriginalBackbone=DeepSeek-R1-Distill-Qwen-14B, Zero-shot Chain-of-Thought (CoT)=true2025.0892.31,917—
DeepSeek-R1-7BMode=Original, Base Model=DeepSeek-R1-7B2025.1092.031,214—