ResearchBenchmarksData Analysis Question Answering on AggQA FinancialFollow95.65Accuracy (Easy)FullContext14.259635.389856.5277.6502Oct 12, 2025Evaluation ResultsMethodMethodLinksAccuracy (Easy)Accuracy (Medium)Accuracy (Hard)Accuracy (Overall)FullContextBackbone=GPT-5 MiniBackbone=GPT-5 Mini2025.1095.6562.557.1470.59PISA2025.1078.2662.578.5769.41LangMem2025.1043.4818.75022.35AMem2025.1039.132542.8630.59Mem02025.1039.1316.6721.4323.53MemoryOS2025.1017.3914.897.1414.29