AfroBench: leaderboard
64 African languages across 15 NLP tasks and 22 datasets (understanding, text generation, knowledge and question answering, math reasoning), from McGill NLP and Masakhane (2025); average score.
Metric: Average Score (%). Source: huggingface.co. Status: saturation imminent. 24 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 | 77.74 |
| 2 | Gemini 3 Pro | 76.01 |
| 3 | Gemini 2.5 Pro | 74.53 |
| 4 | Claude Sonnet 4.5 | 70.6 |
| 5 | Claude Sonnet 4 | 68.7 |
| 6 | Gemini 2.5 Flash | 66.71 |
| 7 | Gemini 2.0 Flash | 66.43 |
| 8 | GPT-4o (2024-08-06) | 65.8 |
| 9 | GPT-4.1 | 65.7 |
| 10 | Gemini 1.5 Pro | 62.8 |
| 11 | Claude 3.7 Sonnet | 60.3 |
| 12 | Gemma 3 27B | 51.01 |
| 13 | Gemma 2 27B | 44.9 |
| 14 | Gemma 2 9B | 40.8 |
| 15 | Llama 3.1 70B | 40 |
Interactive version: theaggregate.ai/benchmark?slug=afrobench · How It Works · Data refreshed daily, snapshot 2026-09-05.