List your company
ResearchFormerly WizWand

Mathematical Reasoning on LiveBench Math

58.1Initial Task ScoreRPT
39.944.62549.3554.075May 20, 2026
Updated 5mo ago

Evaluation results

MethodLinks
RPTOptimizer LLM=GPT-5, Target Model=GPT-4.12026.0558.1——70.5
RPTOptimizer LLM=GPT-5-mini, Target Model=GPT-4.12026.0558.1——66.2
RPTOptimizer LLM=Gemini-3.1-Pro, Target Model=GPT-4.12026.0558.1——70.4
RPTOptimizer LLM=Gemini-3.1-Flash-Lite, Target Model=GPT-4.12026.0558.1——64.9
ACEOptimizer LLM=GPT-5, Target Model=GPT-4.12026.0554.1——61.6
ACEOptimizer LLM=GPT-5-mini, Target Model=GPT-4.12026.0554.1——64.9
ACEOptimizer LLM=Gemini-3.1-Pro, Target Model=GPT-4.12026.0554.1——60.8
ACEOptimizer LLM=Gemini-3.1-Flash-Lite, Target Model=GPT-4.12026.0554.1——56.6
MIPROOptimizer LLM=GPT-5, Target Model=GPT-4.12026.0540.7——47.6
MIPROOptimizer LLM=GPT-5-mini, Target Model=GPT-4.12026.0540.7——43.8
MIPROOptimizer LLM=Gemini-3.1-Pro, Target Model=GPT-4.12026.0540.7——43.2
MIPROOptimizer LLM=Gemini-3.1-Flash-Lite, Target Model=GPT-4.12026.0540.7——42.5
GEPAOptimizer LLM=GPT-5, Target Model=GPT-4.12026.0540.6——46.2
GEPAOptimizer LLM=GPT-5-mini, Target Model=GPT-4.12026.0540.6——46.9
GEPAOptimizer LLM=Gemini-3.1-Pro, Target Model=GPT-4.12026.0540.6——46.9
GEPAOptimizer LLM=Gemini-3.1-Flash-Lite, Target Model=GPT-4.12026.0540.6——42.7
GEPA-COptimizer LLM=GPT-52026.05—40.10.543—
GEPA-COptimizer LLM=GPT-5-mini2026.05—40.10.543—
GEPA-COptimizer LLM=Gemini-3.1-Pro2026.05—40.10.543—
GEPA-COptimizer LLM=Gemini-3.1-FL2026.05—40.10.543—
RPTOptimizer LLM=GPT-52026.05—58.10.347—
RPTOptimizer LLM=GPT-5-mini2026.05—58.10.347—
RPTOptimizer LLM=Gemini-3.1-Pro2026.05—58.10.347—
RPTOptimizer LLM=Gemini-3.1-FL2026.05—58.10.347—