ResearchBenchmarksMathematical Reasoning on GSM8K (Mean/Std Accuracy)Follow95.38Accuracy (mean)MIPRO94.922495.041295.1695.2788May 14, 2026Evaluation ResultsMethodMethodLinksAccuracy (mean)Std Dev AccuracyMIPROModel=Gemini-2-FlashModel=Gemini-2-Flash2026.0595.380.9PSAO.LLMModel=Gemini-2-FlashModel=Gemini-2-Flash2026.0594.950.7COPROModel=Gemini-2-FlashModel=Gemini-2-Flash2026.0594.940.73