SAHARA - Named Entity Recognition: leaderboard
Named entity recognition on texts in African languages, from the SAHARA African-language LLM leaderboard (UBC-NLP, ACL 2025). The task score is the mean over the languages the task covers; Macro-F1 on a 0-100 scale, higher is better.
Metric: Macro-F1 (%). Source: huggingface.co. Saturation forecast: Around February 2027. 48 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 71.06 | #71 |
| 2 | GPT-5 | 66.23 | #105 |
| 3 | Gemini 2.5 Pro | 62.96 | #159 |
| 4 | GPT-5.6 Luna | 59.63 | #79 |
| 5 | GPT-5.1 (2025-11-13) | 58.13 | #107 |
| 6 | Claude Sonnet 4 (20250514) | 35.06 | #223 |
| 7 | Claude Sonnet 4 | 32.89 | #207 |
| 8 | GPT-4.1 | 27.04 | #258 |
| 9 | Qwen 3.6 27B | 24.92 | #182 |
| 10 | Gemini 2.5 Flash | 22.65 | #255 |
| 11 | Qwen 3.5 27B | 21.67 | #192 |
| 12 | Gemma 4 31B (IT) | 20.79 | #178 |
| 13 | c4ai-command-a-03-2025 | 16.88 | #468 |
| 14 | Gemma 4 26B A4B (IT) | 16.46 | #293 |
| 15 | Gemma 2 27B (IT) | 13.14 | #652 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=sahara-named-entity-recognition · How It Works · Data refreshed daily, snapshot 2026-11-05.