AfriEconQA (Closed-Book): leaderboard
AfriEconQA (Ajayi, Alaba and Stephen, 2026) answered from the question alone with no report text, so it measures what a model recalls about 2024 and 2025 World Bank reporting on African economies. Token F1 in percent over the 862 held-out test questions (the paper's zero-shot condition); higher is better.
Metric: Token F1 (%; 862 held-out test questions, question only with no report context, the paper's zero-shot condition). Source: arxiv.org. Saturation forecast: Around January 2028. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek V4 Pro | 26.7 | #181 |
| 2 | Qwen 3.6 35B A3B | 22.8 | #461 |
| 3 | Gemma 4 12B (IT) | 22.3 | #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-closed-book · How It Works · Data refreshed daily, snapshot 2026-11-12.