SAHARA - Topic Classification: leaderboard
Topic classification of 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) | 82.75 | #71 |
| 2 | GPT-5 | 82.74 | #105 |
| 3 | GPT-5.1 (2025-11-13) | 80.18 | #107 |
| 4 | GPT-4.1 | 76.57 | #258 |
| 5 | Gemini 2.5 Pro | 75.65 | #159 |
| 6 | GPT-5.6 Luna | 75.39 | #79 |
| 7 | Claude Sonnet 4 | 74.7 | #207 |
| 8 | Gemma 3 27B (IT) | 70.73 | #528 |
| 9 | Qwen 3.5 27B | 68.72 | #192 |
| 10 | Claude Sonnet 4 (20250514) | 67.93 | #223 |
| 11 | Llama 3.1 70B Instruct | 64.89 | #567 |
| 12 | Qwen 3.6 27B | 64.24 | #182 |
| 13 | Llama 4 Scout Instruct | 59.7 | #547 |
| 14 | Gemini 2.5 Flash | 59.41 | #255 |
| 15 | Llama 3.3 70B Instruct | 56.48 | #541 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=sahara-topic-classification · How It Works · Data refreshed daily, snapshot 2026-11-05.