ResearchTasksFinancial Numerical Reasoning QAFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedTencent Yuanbao financial QA system Online A/B 7-day (test)DSA6.82Hallucination Rate2Jun 1, 2026