List your company
ResearchFormerly WizWand

Math Reasoning on MultiArith (test)

99.59AccuracyZERA
82.024486.584791.14595.7053Apr 8, 2026Apr 9, 2026Apr 10, 2026Apr 11, 2026Apr 12, 2026Apr 13, 2026Apr 14, 2026
Updated 6mo ago

Evaluation results

MethodLinks
ZERAWorker Model=Qwen2.5-7B-Instruct, Architect Model=gpt-oss-120b2026.0499.59
aPSFWorker Model=Qwen2.5-7B-Instruct, Architect Model=gpt-oss-120b2026.0499.53
CriSPOWorker Model=Qwen2.5-7B-Instruct, Architect Model=gpt-oss-120b2026.0499.35
aPSFWorker Model=Qwen2.5-7B-Instruct, Architect Model=Qwen3-8B2026.0499.3
CriSPOWorker Model=Qwen2.5-7B-Instruct, Architect Model=Qwen3-8B2026.0498.94
OPROWorker Model=Qwen2.5-7B-Instruct, Architect Model=Qwen3-8B2026.0498.75
APEWorker Model=Qwen2.5-7B-Instruct, Architect Model=Qwen3-8B2026.0498.54
ProTeGiWorker Model=Qwen2.5-7B-Instruct, Architect Model=Qwen3-8B2026.0498.54
GrIPSWorker Model=Qwen2.5-7B-Instruct, Architect Model=Qwen3-8B2026.0498.54
GrIPSWorker Model=Qwen2.5-7B-Instruct, Architect Model=gpt-oss-120b2026.0498.54
ZERAWorker Model=Qwen2.5-7B-Instruct, Architect Model=Qwen3-8B2026.0498.42
aPSFWorker Model=Llama-3.1-8B-Instruct, Architect Model=Qwen3-8B2026.0498.33
ZERAWorker Model=Llama-3.1-8B-Instruct, Architect Model=Qwen3-8B2026.0498.22
ZERAWorker Model=Llama-3.1-8B-Instruct, Architect Model=gpt-oss-120b2026.0498.15
ProTeGiWorker Model=Qwen2.5-7B-Instruct, Architect Model=gpt-oss-120b2026.0498.12
aPSFWorker Model=Llama-3.1-8B-Instruct, Architect Model=gpt-oss-120b2026.0497.92
CriSPOWorker Model=Llama-3.1-8B-Instruct, Architect Model=gpt-oss-120b2026.0497.9
ProTeGiWorker Model=Llama-3.1-8B-Instruct, Architect Model=gpt-oss-120b2026.0497.5
OPROWorker Model=Qwen2.5-7B-Instruct, Architect Model=gpt-oss-120b2026.0497.29
APEWorker Model=Llama-3.1-8B-Instruct, Architect Model=gpt-oss-120b2026.0496.88
GrIPSWorker Model=Llama-3.1-8B-Instruct, Architect Model=gpt-oss-120b2026.0496.88
APEWorker Model=Llama-3.1-8B-Instruct, Architect Model=Qwen3-8B2026.0496.67
GrIPSWorker Model=Llama-3.1-8B-Instruct, Architect Model=Qwen3-8B2026.0496.67
CriSPOWorker Model=Llama-3.1-8B-Instruct, Architect Model=Qwen3-8B2026.0496.49
Zero-shot CoTWorker Model=Qwen2.5-7B-Instruct, Optimization Protocol=No optimization (p_init)2026.0496.46
OPROWorker Model=Llama-3.1-8B-Instruct, Architect Model=Qwen3-8B2026.0496.46
ProTeGiWorker Model=Llama-3.1-8B-Instruct, Architect Model=Qwen3-8B2026.0496.46
Agent-GWOBackbone=Gemma-3-12b-it2026.0495.9
Agent-GWOBackbone=GPT-4o-mini2026.0495.3
Zero-shot CoTWorker Model=Llama-3.1-8B-Instruct, Optimization Protocol=No optimization (p_init)2026.0495.21
OPROWorker Model=Llama-3.1-8B-Instruct, Architect Model=gpt-oss-120b2026.0495.21
Agent-GWOBackbone=Qwen2.5-Coder-7B-Instruct2026.0493.3
AoTBackbone=GPT-4o-mini2026.0492.6
AFlowBackbone=GPT-4o-mini2026.0491.9
GoTBackbone=GPT-4o-mini2026.0491
Self-RefineBackbone=GPT-4o-mini2026.0489.9
ToTBackbone=GPT-4o-mini2026.0489.8
CoT-SC/n=5Backbone=GPT-4o-mini2026.0489.7
CoTBackbone=GPT-4o-mini2026.0489.5
AFlowBackbone=Gemma-3-12b-it2026.0489
AoTBackbone=Qwen2.5-Coder-7B-Instruct2026.0488.4
ToTBackbone=Gemma-3-12b-it2026.0488.2
Self-RefineBackbone=Gemma-3-12b-it2026.0487.5
AoTBackbone=Gemma-3-12b-it2026.0487.5
CoT-SC/n=5Backbone=Qwen2.5-Coder-7B-Instruct2026.0486.8
AFlowBackbone=Qwen2.5-Coder-7B-Instruct2026.0486.8
GoTBackbone=Gemma-3-12b-it2026.0486.8
APEWorker Model=Qwen2.5-7B-Instruct, Architect Model=gpt-oss-120b2026.0486.46
Self-RefineBackbone=Qwen2.5-Coder-7B-Instruct2026.0486.4
GoTBackbone=Qwen2.5-Coder-7B-Instruct2026.0485.1
CoT-SC/n=5Backbone=Gemma-3-12b-it2026.0485.1
CoTBackbone=Qwen2.5-Coder-7B-Instruct2026.0484.6
ToTBackbone=Qwen2.5-Coder-7B-Instruct2026.0483
CoTBackbone=Gemma-3-12b-it2026.0482.7