[KongchangAI]
BenchmarkFP&A Benchmark / 财务结账基准测试

FP&A Reasoning Benchmark

A financial reasoning benchmark designed by an FP&A practitioner, comprising 28 questions covering tasks such as variance bridging, accrual accounting, financial statement cross-referencing, and assumption validation. It is used to evaluate the reasoning accuracy of large language models in financial close scenarios.

Timeline (last 90 days)

Oct 11

一位拥有12年财务经验的从业者Marco Polo设计了一套FP&A推理基准测试

Unverified50%
Oct 11

该基准由28个小型财务结账任务组成,包括6道差异桥接题、6道权责发生制与截止测试、6道报表勾稽题和10道前提校验题

Unverified50%
Oct 11

评分器要求每个金额与标准答案的误差在£0.005以内,前提标志正确,且桥接表在适用处必须对账成功,哪怕错1便士也判定为失败

Unverified50%
Oct 11

20 of the 41 complete test suites achieved a perfect score of 100%, rendering the benchmark unable to differentiate among these leading models.

Expired50%
Oct 11

Variance bridge questions had the worst failure rate, with a strict pass rate of only 144/246 (58.5%).

Expired50%

All Facts (3)

Source Articles