AfriEconQA - Multiple Choice: leaderboard
The four-option multiple-choice questions of AfriEconQA (Ajayi, Alaba and Stephen, 2026), with distractors from the same economic context in World Bank reports on African economies, answered from the top five chunks of a fixed hybrid BM25 and BGE-M3 retriever over the 862-question held-out test split. F1 in percent on the option letter and short answer phrase; higher is better.
Metric: F1 (%; the four-option multiple-choice questions (answered with the option letter and a short phrase) of the 862-question held-out test split, answered from the top five chunks of the fixed hybrid BM25 + BGE-M3 retriever). Source: arxiv.org. Saturation forecast: Estimated already saturated. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek V4 Pro | 88.4 | #181 |
| 2 | Qwen 3.6 35B A3B | 85.7 | #461 |
| 3 | Gemma 4 12B (IT) | 78.3 | #688 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=afrieconqa-multiple-choice · How It Works · Data refreshed daily, snapshot 2026-11-12.