List your company

Grade School Math Reasoning on GSM8K (Accuracy)

98Accuracy (GSM8K)PEAR
8588.37591.7595.125Jun 4, 2025Aug 2, 2025Sep 30, 2025Nov 29, 2025Jan 27, 2026Mar 27, 2026May 26, 2026

Evaluation results

MethodLinks
PEARBackbone=Qwen-3-30B-A3B2026.0598
CliqueBackbone=Qwen-3-30B-A3B2026.0596
S-BoN (b)n=2562025.0695.6
S-BoN (b)n=642025.0695.5
PEARBackbone=Gemma-3-12B2026.0595.5
StarBackbone=Qwen-3-30B-A3B2026.0595.5
RandomBackbone=Qwen-3-30B-A3B2026.0595.5
PEARBackbone=Claude-Haiku-4.52026.0595.5
AoTBackbone=DeepSeek-V32026.0395.1
AoTBackbone=GPT-4o-mini2026.0395
ToTBackbone=DeepSeek-V32026.0395
RingBackbone=Qwen-3-30B-A3B2026.0595
PEARBackbone=GPT-5.4-nano2026.0595
ToTBackbone=GPT-4o-mini2026.0394.9
AoTBackbone=Qwen-Turbo2026.0394.7
S-BoN (b)n=42025.0694.7
Re2Backbone=Qwen2.5-14B Base, Model Category=Base Model2026.0394.6
GoTBackbone=GPT-4o-mini2026.0394.5
GoTBackbone=DeepSeek-V32026.0394.5
S-BoN (b)n=162025.0694.5
PEARBackbone=Qwen-2.5-14B2026.0594.5
ChainBackbone=Qwen-3-30B-A3B2026.0594.5
TDA-RCBackbone=DeepSeek-V32026.0394.4
DAPOBackbone=Qwen2.5-14B Base, Model Category=Base Model2026.0394.3
GSIn=2562025.0694.3
TDA-RCBackbone=GPT-4o-mini2026.0394.2
ToTBackbone=Qwen-Turbo2026.0394.2
FoT (n=8)Backbone=DeepSeek-V32026.0394.2
Re2Backbone=Qwen2.5-7B-Instruct, Model Category=Instruct Model2026.0394.1
FoT (n=8)Backbone=GPT-4o-mini2026.0394
FoT (n=8)Backbone=Qwen-Turbo2026.0393.9
GoTBackbone=Qwen-Turbo2026.0393.8
TDA-RCBackbone=Qwen-Turbo2026.0393.7
Re2Backbone=Qwen2.5-7B Base, Model Category=Base Model2026.0393.6
AFlowBackbone=DeepSeek-V32026.0393.6
AFlowBackbone=GPT-4o-mini2026.0393.5
GSIn=642025.0693.5
S-BoN (b)n=12025.0693.3
AFlowBackbone=Qwen-Turbo2026.0393.1
ChainBackbone=Qwen-2.5-14B2026.0593
GSIn=162025.0692.8
DAPOBackbone=Qwen2.5-7B-Instruct, Model Category=Instruct Model2026.0392.6
GSIn=42025.0692.4
Re2Backbone=DeepSeek-R1-Distill-Llama-8B, Model Category=Reasoning Model2026.0392.2
CoT-SCBackbone=DeepSeek-V32026.0392.1
CoT-SC (n=5)Backbone=GPT-4o-mini2026.0392
Self-RefineBackbone=DeepSeek-V32026.0391.9
DAPOBackbone=Qwen2.5-7B Base, Model Category=Base Model2026.0391.8
CoFiCotBase Model=Llama3-8B-Instruct2026.0391.8
Self-RefineBackbone=GPT-4o-mini2026.0391.7
RSDn=2562025.0691.6
CoT-SC (n=5)Backbone=Qwen-Turbo2026.0391.5
StarBackbone=Gemma-3-12B2026.0591.5
RingBackbone=Qwen-2.5-14B2026.0591.5
RandomBackbone=Claude-Haiku-4.52026.0591.5
Self-RefineBackbone=Qwen-Turbo2026.0391.4
CoTBackbone=DeepSeek-V32026.0391.3
CoFiCotBase Model=GPT-3.5-Turbo2026.0391.2
RSDn=162025.0691
CliqueBackbone=Claude-Haiku-4.52026.0591
CoTBackbone=GPT-4o-mini2026.0390.9
RSDn=642025.0690.9
CoTBackbone=Qwen-Turbo2026.0390.7
Best-of-kBase Model=GPT-3.5-Turbo, k=1202026.0390.5
RandomBackbone=Gemma-3-12B2026.0590.5
RandomBackbone=Qwen-2.5-14B2026.0590.5
CliqueBackbone=GPT-5.4-nano2026.0590.5
ChainBackbone=Claude-Haiku-4.52026.0590.5
k-way SCBase Model=Llama3-8B-Instruct, k=1202026.0390.3
Best-of-kBase Model=Llama3-8B-Instruct, k=1202026.0390.1
Self-Refine + k-way SCBase Model=Llama3-8B-Instruct, k=1202026.0390.1
CliqueBackbone=Gemma-3-12B2026.0590
RandomBackbone=GPT-5.4-nano2026.0590
S-BoN (s)n=2562025.0689.9
k-way SCBase Model=GPT-3.5-Turbo, k=1202026.0389.8
DARModel=Falcon3-7B2026.0389.7
DAPOBackbone=DeepSeek-R1-Distill-Llama-8B, Model Category=Reasoning Model2026.0389.6
S-BoN (s)n=642025.0689.6
RSDn=42025.0689.5
PEARBackbone=Llama-3.1-8B2026.0589.5
StarBackbone=Qwen-2.5-14B2026.0589.5
StarBackbone=GPT-5.4-nano2026.0589.5
RingBackbone=Claude-Haiku-4.52026.0589.5
RingBackbone=Gemma-3-12B2026.0589
CliqueBackbone=Qwen-2.5-14B2026.0589
StarBackbone=Claude-Haiku-4.52026.0589
ChainBackbone=Gemma-3-12B2026.0588.5
ChainBackbone=GPT-5.4-nano2026.0588.5
GSIn=12025.0688.3
Self-Refine + k-way SCBase Model=GPT-3.5-Turbo, k=1202026.0388.1
S-BoN (s)n=162025.0688.1
Majority VoteModel=Falcon3-7B2026.0388
RingBackbone=GPT-5.4-nano2026.0588
ChainBackbone=Llama-3.1-8B2026.0587.5
Vote PromptModel=Falcon3-7B2026.0387
Uncertain PromptModel=Falcon3-7B2026.0386.7
StarBackbone=Llama-3.1-8B2026.0586.5
DARModel=Llama3.1-8B2026.0386.3
RandomBackbone=Llama-3.1-8B2026.0586
Society Of MindModel=Llama3.1-8B2026.0385.5