List your company

Mathematical Reasoning on LiveBench Math (test)

51.95ScoreGEPA
46.38647.830549.27550.7195Jul 25, 2025

Evaluation results

MethodLinks
GEPABackbone=Qwen3 8B, Optimization Budget (# rollouts)=18392025.0751.95
GEPA+MergeBackbone=Qwen3 8B, Optimization Budget (# rollouts)=18392025.0751.95
GRPOBackbone=Qwen3 8B, Optimization Budget (# rollouts)=240002025.0751.26
BaselineBackbone=Qwen3 8B2025.0748.7
MIPROv2Backbone=Qwen3 8B2025.0746.6