ResearchBenchmarksPrompt Optimization on PUPAFollow91.85ScoreGEPA80.378883.356986.33589.3131May 10, 2026Evaluation ResultsMethodMethodLinksScoreOptimization Budget (# Rollouts)GEPATask Model=Qwen 3 8BTask Model=Qwen 3 8B2026.0591.852,426LEVITask Model=Qwen 3 8BTask Model=Qwen 3 8B2026.0589.731,275MIPROv2Task Model=Qwen 3 8BTask Model=Qwen 3 8B2026.0581.55—BaselineTask Model=Qwen 3 8BTask Model=Qwen 3 8B2026.0580.82—