AfriEconQA Benchmark for Quantitative and Temporal Reasoning
July 31, 2026
AfriEconQA introduces a benchmark of 4,309 evidence-linked QA instances derived from 220 World Bank economic reports. The dataset specifically tests an LLM's ability to perform precise numerical extraction and temporal reasoning across dense, repetitive institutional prose.
HOW THIS AFFECTS YOU
●
researcherYou can use this to evaluate how well your models handle long-document quantitative and temporal reasoning tasks.