List your company
ResearchFormerly WizWand

Math Reasoning on MATH (test) (Accuracy)

96AccuracyMAD
-1.2424.00549.2574.495May 17, 2025Jul 15, 2025Sep 13, 2025Nov 11, 2025Jan 10, 2026Mar 10, 2026May 9, 2026
Updated 5mo ago

Evaluation results

MethodLinks
MADCategory=Multi-Agent2026.0596
MoACategory=Multi-Agent2026.0595.6
DMADCategory=Multi-Agent2026.0595
AgentPSOCategory=Multi-Agent2026.0594.6
Self-RefineCategory=Advanced Single-Agent2026.0591.6
ReflectionCategory=Advanced Single-Agent2026.0590.8
Chain-of-ThoughtCategory=Vanilla Single-Agent2026.0590.4
Step-Back PromptingCategory=Vanilla Single-Agent2026.0590.4
Self-ConsistencyCategory=Advanced Single-Agent2026.0588.8
AoTBackbone=GPT-4o-mini2026.0483.6
Agent-GWOBackbone=Gemma-3-12b-it2026.0482.1
ToTCategory=Advanced Single-Agent2026.0581.6
GoTBackbone=Gemma-3-12b-it2026.0480.3
Agent-GWOBackbone=GPT-4o-mini2026.0480.2
AoTBackbone=Gemma-3-12b-it2026.0479.6
ToTBackbone=Gemma-3-12b-it2026.0479.5
AFlowBackbone=Gemma-3-12b-it2026.0479
AFlowBackbone=GPT-4o-mini2026.0478.9
GoTBackbone=GPT-4o-mini2026.0478.6
ToTBackbone=GPT-4o-mini2026.0477.8
Self-RefineBackbone=Gemma-3-12b-it2026.0477.6
CoT-SC/n=5Backbone=GPT-4o-mini2026.0476.3
Self-RefineBackbone=GPT-4o-mini2026.0475.1
CoTBackbone=GPT-4o-mini2026.0474.8
CoT-SC/n=5Backbone=Gemma-3-12b-it2026.0474.6
Agent-GWOBackbone=Qwen2.5-Coder-7B-Instruct2026.0474.1
CoTBackbone=Gemma-3-12b-it2026.0472.8
AoTBackbone=Qwen2.5-Coder-7B-Instruct2026.0472
AFlowBackbone=Qwen2.5-Coder-7B-Instruct2026.0471.9
GoTBackbone=Qwen2.5-Coder-7B-Instruct2026.0471.7
ToTBackbone=Qwen2.5-Coder-7B-Instruct2026.0471.3
Self-RefineBackbone=Qwen2.5-Coder-7B-Instruct2026.0470.4
CoT-SC/n=5Backbone=Qwen2.5-Coder-7B-Instruct2026.0467.3
CoTBackbone=Qwen2.5-Coder-7B-Instruct2026.0465.8
MENTORCOLLAB MLPGenerator=Qwen3-8B-Base, Mentor=Qwen3-32B, rho=25%, Decoding Strategy=greedy, Token Budget=5122026.0246.8
OMACOptimization Dimension Category=Functional Dimension, Specific Dimension Optimized=Fun-1.12025.0535.17
OMACOptimization Dimension Category=Functional Dimension, Specific Dimension Optimized=Fun-1.22025.0534.91
OMACOptimization Dimension Category=Functional Dimension, Specific Dimension Optimized=Fun-22025.0533.95
OMACOptimization Dimension Category=Structural Dimension, Specific Dimension Optimized=Str-32025.0533.7
OMACOptimization Dimension Category=Structural Dimension, Specific Dimension Optimized=Str-22025.0533.41
OMACOptimization Dimension Category=Structural Dimension, Specific Dimension Optimized=Str-12025.0533.34
AFlowMethod Category=Baselines2025.0532.49
DyLANMethod Category=Baselines2025.0532.35
LLM DebateMethod Category=Baselines2025.0529.42
ADASMethod Category=Baselines2025.0528.94
SEMethod Category=Baselines2025.0528.72
Co-LLMGenerator=Qwen3-8B-Base, Mentor=Qwen3-14B2026.0225
MENTORCOLLAB MLPGenerator=Gemma-3-4B-PT, Mentor=Qwen3-14B, rho=25%, Decoding Strategy=greedy, Token Budget=5122026.0221
MENTORCOLLAB MLPGenerator=Llama3.1-8B, Mentor=Qwen3-32B, rho=25%, Decoding Strategy=greedy, Token Budget=5122026.0218
MENTORCOLLAB FREEGenerator=Gemma-3-4B-PT, Mentor=Qwen3-14B, rho=25%, Decoding Strategy=greedy, Token Budget=5122026.0215.8
Generator BaselineModel=Gemma-3-4B-PT, Decoding Strategy=greedy, Token Budget=5122026.0214.2
NudgingGenerator=Qwen3-1.7B, Mentor=R1-Distilled-Llama-70B, gamma=0.4, Decoding Strategy=greedy, Token Budget=5122026.0213.6
FedMomentumBackbone=LLaMA2-7B, LoRA rank=32, scaling factor=64, communication rounds=502026.034.44
RoLoRABackbone=LLaMA2-7B, LoRA rank=32, scaling factor=64, communication rounds=502026.034.42
FFA-LoRABackbone=LLaMA2-7B, LoRA rank=32, scaling factor=64, communication rounds=502026.034.18
FedEx-LoRABackbone=LLaMA2-7B, LoRA rank=32, scaling factor=64, communication rounds=502026.033.97
FLoRABackbone=LLaMA2-7B, LoRA rank=32, scaling factor=64, communication rounds=502026.033.86
FedITBackbone=LLaMA2-7B, LoRA rank=32, scaling factor=64, communication rounds=502026.033.84
Pre-trainedBackbone=LLaMA2-7B2026.032.5