List your company
ResearchFormerly WizWand

Math Reasoning on GSM-Hard (test)

55.94AccuracyCriSPO
30.075236.790143.50550.2199Apr 8, 2026
Updated 6mo ago

Evaluation results

MethodLinks
CriSPOWorker Model=Qwen2.5-7B-Instruct, Architect Model=Qwen3-8B2026.0455.94
aPSFWorker Model=Qwen2.5-7B-Instruct, Architect Model=gpt-oss-120b2026.0455.86
ProTeGiWorker Model=Qwen2.5-7B-Instruct, Architect Model=gpt-oss-120b2026.0455.59
aPSFWorker Model=Qwen2.5-7B-Instruct, Architect Model=Qwen3-8B2026.0454.77
APEWorker Model=Qwen2.5-7B-Instruct, Architect Model=gpt-oss-120b2026.0454.55
ProTeGiWorker Model=Qwen2.5-7B-Instruct, Architect Model=Qwen3-8B2026.0453.88
Zero-shot CoTWorker Model=Qwen2.5-7B-Instruct, Optimization Protocol=No optimization (p_init)2026.0453.79
OPROWorker Model=Qwen2.5-7B-Instruct, Architect Model=Qwen3-8B2026.0453.6
GrIPSWorker Model=Qwen2.5-7B-Instruct, Architect Model=gpt-oss-120b2026.0453.41
ZERAWorker Model=Qwen2.5-7B-Instruct, Architect Model=Qwen3-8B2026.0453.22
ZERAWorker Model=Qwen2.5-7B-Instruct, Architect Model=gpt-oss-120b2026.0452.63
GrIPSWorker Model=Qwen2.5-7B-Instruct, Architect Model=Qwen3-8B2026.0452.56
APEWorker Model=Qwen2.5-7B-Instruct, Architect Model=Qwen3-8B2026.0451.7
OPROWorker Model=Qwen2.5-7B-Instruct, Architect Model=gpt-oss-120b2026.0451.69
CriSPOWorker Model=Qwen2.5-7B-Instruct, Architect Model=gpt-oss-120b2026.0451.25
aPSFWorker Model=Llama-3.1-8B-Instruct, Architect Model=gpt-oss-120b2026.0442.96
aPSFWorker Model=Llama-3.1-8B-Instruct, Architect Model=Qwen3-8B2026.0435.98
ProTeGiWorker Model=Llama-3.1-8B-Instruct, Architect Model=gpt-oss-120b2026.0435.88
CriSPOWorker Model=Llama-3.1-8B-Instruct, Architect Model=gpt-oss-120b2026.0435.16
CriSPOWorker Model=Llama-3.1-8B-Instruct, Architect Model=Qwen3-8B2026.0434.94
ZERAWorker Model=Llama-3.1-8B-Instruct, Architect Model=gpt-oss-120b2026.0434.7
ZERAWorker Model=Llama-3.1-8B-Instruct, Architect Model=Qwen3-8B2026.0434.58
OPROWorker Model=Llama-3.1-8B-Instruct, Architect Model=gpt-oss-120b2026.0434.47
OPROWorker Model=Llama-3.1-8B-Instruct, Architect Model=Qwen3-8B2026.0433.62
Zero-shot CoTWorker Model=Llama-3.1-8B-Instruct, Optimization Protocol=No optimization (p_init)2026.0433.6
ProTeGiWorker Model=Llama-3.1-8B-Instruct, Architect Model=Qwen3-8B2026.0433.52
APEWorker Model=Llama-3.1-8B-Instruct, Architect Model=Qwen3-8B2026.0433.05
GrIPSWorker Model=Llama-3.1-8B-Instruct, Architect Model=gpt-oss-120b2026.0432.95
GrIPSWorker Model=Llama-3.1-8B-Instruct, Architect Model=Qwen3-8B2026.0432.86
APEWorker Model=Llama-3.1-8B-Instruct, Architect Model=gpt-oss-120b2026.0431.07